{"name":"AI Wiki open-weight model directory","description":"Models with downloadable weights covered by AI Wiki, with the license of the weights, whether commercial use is allowed, parameter count and release date. Derived from the fact box of each model article; the article is the source for every value.","license":"CC BY 4.0","license_url":"https://creativecommons.org/licenses/by/4.0/","attribution":"AI Wiki (aiwiki.ai)","homepage":"https://aiwiki.ai/tools/open_models","generated_at":"2026-11-17T14:57:23.331Z","tiers":{"permissive":"Weights are downloadable under a license that allows any use, including commercial use, modification and redistribution.","conditional":"Weights are downloadable under the developer's own terms. Commercial use is allowed, with conditions such as an acceptable-use policy, a revenue or user threshold, or registration.","noncommercial":"Weights are downloadable for research or personal use. Commercial use needs a separate agreement.","partial":"Some models in this family have downloadable weights. Others are available only through an API or a product.","closed":"Weights are not published. The model is available only through an API or a product."},"model_count":410,"models":[{"name":"ERNIE 5.5 Flash Base","article":"https://aiwiki.ai/wiki/ernie_5_5_flash_base","kind":"language","developer":"Baidu","released":"2026-11-05","parameters":"280B total, 14B active","parameters_billions":280,"tier":"conditional","weights":"open","license":"custom","license_name":"ERNIE 5.5 Flash Base Model License Agreement v1.0","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Excludes the EU, UK and South Korea; commercial use needs notice and attribution up to US$50M revenue and written authorization above it.","hugging_face":"https://huggingface.co/baidu/ERNIE-5.5-Flash-Base","article_fact_checked":false},{"name":"Hydrogen-1.0","article":"https://aiwiki.ai/wiki/hydrogen_1_0","kind":"science","developer":"Deep Principle","released":"2026-11-02","parameters":"About 753B (753,329,940,480 in the BF16 checkpoint)","parameters_billions":753,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":false,"note":"The Hugging Face repository is gated: users accept access conditions before download.","hugging_face":"https://huggingface.co/deep-principle/Hydrogen-1.0","article_fact_checked":false},{"name":"LongCat-2.5","article":"https://aiwiki.ai/wiki/longcat_2_5","kind":"language","developer":"Meituan LongCat team","released":"2026-11-02","parameters":"1.6T total, about 48B active","parameters_billions":1600,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"The hosted launch came first; the weights were published on Hugging Face in November 2026.","hugging_face":"https://huggingface.co/meituan-longcat/LongCat-2.5","article_fact_checked":false},{"name":"Ripple-65B Preview","article":"https://aiwiki.ai/wiki/ripple_65b_preview","kind":"language","developer":"Ricursive Intelligence","released":"2026-11","parameters":"About 65B","parameters_billions":65,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/ricursiveAI/Ripple-65B-Preview","article_fact_checked":false},{"name":"Pika H3 Flex","article":"https://aiwiki.ai/wiki/pika_h3_flex","kind":"video","developer":"Pika (Mellis, Inc.)","released":"2026-10-29","parameters":"33B base generator (MiniMax H3) with two rank-128 LoRA adapters","parameters_billions":33,"tier":"conditional","weights":"open","license":"custom","license_name":"MiniMax H3 Community License Agreement","commercial_use":"conditional","training_code_published":null,"training_data_published":false,"note":"Excludes the EU, UK, South Korea and US; commercial use above US$20M annual revenue needs MiniMax's written authorization.","hugging_face":"https://huggingface.co/Pika-Labs/Pika-H3-Flex","article_fact_checked":true},{"name":"Nemotron 3.5 Transcribe","article":"https://aiwiki.ai/wiki/nemotron_3_5_transcribe","kind":"audio","developer":"NVIDIA","released":"2026-10-28","parameters":"About 34B total, about 4.3B active (including the audio encoder)","parameters_billions":34,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Transcribe-30B-A4B","article_fact_checked":true},{"name":"LFM3","article":"https://aiwiki.ai/wiki/lfm3","kind":"language","developer":"Liquid AI","released":"2026-10-27","parameters":"1.73B to 2.17B (six post-trained models and the LFM3-1.7B-Base checkpoint)","parameters_billions":1.73,"tier":"conditional","weights":"open","license":"custom","license_name":"LFM Open License v1.0","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Commercial use is licensed only for entities under $10M annual revenue; qualified non-profits doing research are exempt from the threshold.","hugging_face":null,"article_fact_checked":true},{"name":"Kimi K3.1 Pro","article":"https://aiwiki.ai/wiki/kimi_k3_1_pro","kind":"language","developer":"Moonshot AI","released":"2026-10-23","parameters":"2.8T total, 104B active","parameters_billions":2800,"tier":"conditional","weights":"open","license":"custom","license_name":"Kimi K3.1-Pro License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Model-as-a-service firms over $20M yearly revenue need a separate agreement; very large products must display Kimi K3.1-Pro.","hugging_face":"https://huggingface.co/moonshotai/Kimi-K3.1-Pro","article_fact_checked":true},{"name":"Nemotron 3.5 Bolt E4B","article":"https://aiwiki.ai/wiki/nemotron_3_5_bolt_e4b","kind":"language","developer":"NVIDIA","released":"2026-10-23","parameters":"8B total (4B effective)","parameters_billions":8,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Post-trained from Google's Gemma 4 E4B, whose own weights are under Apache 2.0.","hugging_face":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-E4B-4GB","article_fact_checked":true},{"name":"WeVisOne-VL","article":"https://aiwiki.ai/wiki/wevisone_vl","kind":"vision","developer":"Tencent (WeChat Vision team)","released":"2026-10-22","parameters":"2B and 4B variants, each with a 9.6M pixel decoder","parameters_billions":2,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/tencent/WeVisOne-VL-4B","article_fact_checked":false},{"name":"AW-0","article":"https://aiwiki.ai/wiki/aw_0","kind":"robotics","developer":"Researchers at UC Berkeley, NVIDIA and the University of Pennsylvania","released":"2026-10-21","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"bsd-2-clause","license_name":"BSD 2-Clause","commercial_use":"yes","training_code_published":true,"training_data_published":false,"note":null,"hugging_face":"https://huggingface.co/miniFranka/agentic-wam-pretrain-yamego3k-soup-e1","article_fact_checked":false},{"name":"Recursive Flint","article":"https://aiwiki.ai/wiki/recursive_flint","kind":"language","developer":"Recursive Superintelligence","released":"2026-10-21","parameters":"39.8B total; approximately 2.5B active per token","parameters_billions":39.8,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/RecursiveSI/flint","article_fact_checked":false},{"name":"Spark-X2.6-4B","article":"https://aiwiki.ai/wiki/spark_x2_6_4b","kind":"language","developer":"SparkLLM / XHToken","released":"2026-10-21","parameters":"4,112,079,360","parameters_billions":4.11207936,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/XHToken/Spark-X2.6-4B","article_fact_checked":false},{"name":"Ornith-1.5-122B","article":"https://aiwiki.ai/wiki/ornith_1_5_122b","kind":"language","developer":"Ornith (ornith-ai)","released":"2026-10-20","parameters":"125.1B in checkpoint metadata (named 122B)","parameters_billions":125.1,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/ornith-ai/Ornith-1.5-122B","article_fact_checked":false},{"name":"MiniMax M3.1 Flash","article":"https://aiwiki.ai/wiki/minimax_m3_1_flash","kind":"language","developer":"MiniMax","released":"2026-10-19","parameters":"About 428B total, 23B active per token","parameters_billions":428,"tier":"conditional","weights":"open","license":"custom","license_name":"MiniMax Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Commercial use needs a Built with MiniMax M3.1-Flash notice and a notice to MiniMax; written authorization above $20M yearly revenue.","hugging_face":"https://huggingface.co/MiniMaxAI/MiniMax-M3.1-Flash","article_fact_checked":true},{"name":"Beam (Reflection)","article":"https://aiwiki.ai/wiki/beam_reflection","kind":"language","developer":"Reflection AI","released":"2026-10-16","parameters":"501.3B total, 23.4B active","parameters_billions":501.3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/reflection/Beam-1.0-501B-A23B","article_fact_checked":true},{"name":"Muse Spark Open","article":"https://aiwiki.ai/wiki/muse_spark_open","kind":"language","developer":"Meta Superintelligence Labs","released":"2026-10-15","parameters":"816B total, 42B active","parameters_billions":816,"tier":"conditional","weights":"open","license":"custom","license_name":"Muse Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated repository. Products must show Built with Muse Spark Open; API providers above $20M yearly revenue need an agreement with Meta.","hugging_face":"https://huggingface.co/meta-models/Muse-Spark-Open-816B-A42B","article_fact_checked":true},{"name":"Nemotron 3.5 Bolt","article":"https://aiwiki.ai/wiki/nemotron_3_5_bolt","kind":"language","developer":"NVIDIA","released":"2026-10-15","parameters":"74B total, 9B active","parameters_billions":74,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16","article_fact_checked":true},{"name":"Nemotron 3.5 Super VL","article":"https://aiwiki.ai/wiki/nemotron_3_5_super_vl","kind":"language","developer":"NVIDIA","released":"2026-10-15","parameters":"120B total, 12B active per token","parameters_billions":120,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16","article_fact_checked":false},{"name":"Qwen-Image-2.1-Turbo","article":"https://aiwiki.ai/wiki/qwen_image_2_1_turbo","kind":"image","developer":"Qwen team (Alibaba)","released":"2026-10-09","parameters":"7B (visual generation component)","parameters_billions":7,"tier":"noncommercial","weights":"open","license":"qwen-research","license_name":"Qwen Research License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Research and evaluation use only; commercial use needs a separate license from Qwen.","hugging_face":"https://huggingface.co/Qwen/Qwen-Image-2.1-Turbo","article_fact_checked":true},{"name":"Underdog Saluki 27B","article":"https://aiwiki.ai/wiki/underdog_saluki_27b","kind":"language","developer":"Underdog (Conway Research)","released":"2026-10-07","parameters":"27B (2-bit quantized; 7.89 GB GGUF file)","parameters_billions":27,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Relicensed from Apache 2.0 to CC BY-NC 4.0 on November 5, 2026; commercial use needs a separate license from Conway Research.","hugging_face":"https://huggingface.co/ConwayResearch/Underdog-Saluki-27B-1.0","article_fact_checked":true},{"name":"EmbeddingGemma 2","article":"https://aiwiki.ai/wiki/embeddinggemma_2","kind":"language","developer":"Google DeepMind","released":"2026-10-06","parameters":"About 740M (all modality encoders loaded)","parameters_billions":0.74,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/google/embeddinggemma-2","article_fact_checked":true},{"name":"Mistral Large 4","article":"https://aiwiki.ai/wiki/mistral_large_4","kind":"language","developer":"Mistral AI","released":"2026-10-06","parameters":"About 1T total, 49B active per token (52B counting embeddings and output layers)","parameters_billions":1000,"tier":"conditional","weights":"open","license":"custom","license_name":"Mistral Ecosystem Open-Weight License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Commercial use allowed, but Model-as-a-Service operators need a separate commercial license from Mistral.","hugging_face":"https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B","article_fact_checked":true},{"name":"Decision 2.0","article":"https://aiwiki.ai/wiki/decision_2_0","kind":"other","developer":"vLLM Semantic Router project","released":"2026-10-03","parameters":"0.60B to 29.37B (six checkpoints, nominal sizes 0.6B to 27B)","parameters_billions":0.6,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":true},{"name":"Kolibri 1","article":"https://aiwiki.ai/wiki/kolibri_1","kind":"language","developer":"Aleph Alpha","released":"2026-10-03","parameters":"78.1B total, 3.46B active (excluding embeddings)","parameters_billions":78.1,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Aleph-Alpha/Kolibri-1","article_fact_checked":true},{"name":"Whistle (Cactus Compute)","article":"https://aiwiki.ai/wiki/whistle_cactus_compute","kind":"audio","developer":"Cactus Compute","released":"2026-10-02","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Cactus-Compute/whistle","article_fact_checked":true},{"name":"SenseNova 6.8 Code2Video","article":"https://aiwiki.ai/wiki/sensenova_6_8_code2video","kind":"video","developer":"SenseTime (SenseNova team)","released":"2026-10","parameters":"27B (27,356,728,560)","parameters_billions":27,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/sensenova/SenseNova-6.8-Code2Video","article_fact_checked":true},{"name":"Intern-Decision","article":"https://aiwiki.ai/wiki/intern_decision","kind":"vision","developer":"InternLM team, Shanghai AI Laboratory","released":"2026-09-26","parameters":"0.8B, 2B and 4B (about 0.85B, 2.21B and 4.54B in the published safetensors)","parameters_billions":0.8,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":false,"note":"Three sizes, all Apache 2.0; the upstream Qwen3.5 license file (also Apache 2.0) is kept alongside.","hugging_face":"https://huggingface.co/internlm/Intern-Decision-4B","article_fact_checked":true},{"name":"Ming-Image-0.1-Design","article":"https://aiwiki.ai/wiki/ming_image_0_1_design","kind":"image","developer":"inclusionAI (Ant Group)","released":"2026-09-22","parameters":"6B diffusion transformer per model (developer figure); the checkpoint also holds a 17B multimodal language model and a 1.5B connector","parameters_billions":6,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/inclusionAI/Ming-Image-0.1-Design","article_fact_checked":true},{"name":"Alice AI Foundation 80B-A3B","article":"https://aiwiki.ai/wiki/alice_ai_foundation_80b_a3b","kind":"language","developer":"Yandex","released":"2026-09-21","parameters":"80B total, 3B active","parameters_billions":80,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/yandex/AliceAI-Foundation-80B-A3B-Base","article_fact_checked":true},{"name":"Xiaomi MiMo-V2.6","article":"https://aiwiki.ai/wiki/mimo_v2_6","kind":"language","developer":"Xiaomi (MiMo team)","released":"2026-09-21","parameters":"310B total, 15B active (Flash); 1.02T total, 42B active (Pro)","parameters_billions":310,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Pro and Flash each ship as RL and MOPD checkpoints, all MIT. RL environments and training code were released too.","hugging_face":null,"article_fact_checked":true},{"name":"Qwen-Image-2.1","article":"https://aiwiki.ai/wiki/qwen_image_2_1","kind":"image","developer":"Qwen team, Alibaba","released":"2026-09-20","parameters":"7B diffusion transformer; 8B Qwen3-VL text encoder","parameters_billions":7,"tier":"noncommercial","weights":"open","license":"qwen-research","license_name":"Qwen Research License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Research and evaluation use only; commercial use needs a separate license from Qwen.","hugging_face":"https://huggingface.co/Qwen/Qwen-Image-2.1","article_fact_checked":true},{"name":"Step 5 Preview","article":"https://aiwiki.ai/wiki/step_5_preview","kind":"language","developer":"StepFun","released":"2026-09-20","parameters":"600B total, 27B active","parameters_billions":600,"tier":"conditional","weights":"open","license":"custom","license_name":"StepFun Model License 1.0","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Device makers and MaaS operators above US$20M yearly revenue need a separate commercial license; large services must show attribution.","hugging_face":"https://huggingface.co/stepfun-ai/Step-5-Preview","article_fact_checked":true},{"name":"DeepSeek V4.1-Flash","article":"https://aiwiki.ai/wiki/deepseek_v4_1_flash","kind":"language","developer":"DeepSeek","released":"2026-09-10","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash","article_fact_checked":true},{"name":"MiniCPM5-2B","article":"https://aiwiki.ai/wiki/minicpm5_2b","kind":"language","developer":"OpenBMB","released":"2026-09-07","parameters":"2.5B (2,516,756,480 total; 1,981,982,720 non-embedding)","parameters_billions":2.5,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":true,"note":"The datasets named on the model card (UltraData family) are published on Hugging Face under Apache 2.0.","hugging_face":"https://huggingface.co/openbmb/MiniCPM5-2B","article_fact_checked":true},{"name":"Extropic Z1T","article":"https://aiwiki.ai/wiki/extropic_z1t","kind":"other","developer":"Extropic","released":"2026-09-04","parameters":null,"parameters_billions":null,"tier":"conditional","weights":"open","license":"unspecified","license_name":"License not stated","commercial_use":null,"training_code_published":true,"training_data_published":null,"note":"Weights repo Extropic-AI/Z1T-0 states no license; the code repo extropic-ai/sparse-transformers is Apache 2.0.","hugging_face":"https://huggingface.co/Extropic-AI/Z1T-0","article_fact_checked":true},{"name":"Ling-3.0-flash-Sante","article":"https://aiwiki.ai/wiki/ling_3_0_flash_sante","kind":"science","developer":"inclusionAI","released":"2026-09-04","parameters":"124B total, 5.1B active per token","parameters_billions":124,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Weights were published on Hugging Face and ModelScope after the September 2026 API-only launch. Research model, not a medical device.","hugging_face":"https://huggingface.co/inclusionAI/Ling-3.0-flash-Sante","article_fact_checked":true},{"name":"Hy4 Preview","article":"https://aiwiki.ai/wiki/hy4_preview","kind":"language","developer":"Tencent Hy Team","released":"2026-08-28","parameters":"770B total, 49B active (backbone)","parameters_billions":770,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":false,"training_data_published":false,"note":null,"hugging_face":"https://huggingface.co/tencent/Hy4-preview","article_fact_checked":true},{"name":"GLM-5.3-Flash","article":"https://aiwiki.ai/wiki/glm_5_3_flash","kind":"language","developer":"Z.ai","released":"2026-08-26","parameters":"320B total, 18B active","parameters_billions":320,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":false,"training_data_published":false,"note":null,"hugging_face":"https://huggingface.co/zai-org/GLM-5.3-Flash","article_fact_checked":true},{"name":"Qwen3.8-Flash-Next","article":"https://aiwiki.ai/wiki/qwen3_8_flash_next","kind":"language","developer":"Qwen team, Alibaba","released":"2026-08-26","parameters":"125B main model, 6B active; plus a 51B N-gram table and a 4B multi-token-prediction module","parameters_billions":125,"tier":"conditional","weights":"open","license":"custom","license_name":"Qwen Community License 1.0","commercial_use":"conditional","training_code_published":false,"training_data_published":false,"note":"Model-as-a-service and AI work-assistant businesses need a separate Qwen license; large commercial products must display the model name.","hugging_face":"https://huggingface.co/Qwen/Qwen3.8-Flash-Next","article_fact_checked":true},{"name":"WeMM-Embedding","article":"https://aiwiki.ai/wiki/wemm_embedding","kind":"vision","developer":"Tencent (WeChat Vision team)","released":"2026-08-25","parameters":"About 9.41B (9,407,831,280 in the published checkpoint)","parameters_billions":9.41,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/tencent/WeMM-Embedding-9B","article_fact_checked":true},{"name":"GLM-5.3","article":"https://aiwiki.ai/wiki/glm_5_3","kind":"language","developer":"Z.ai","released":"2026-08-14","parameters":"744B total, 40B active (Z.ai model-index label)","parameters_billions":744,"tier":"conditional","weights":"open","license":"custom","license_name":"GLM-5.3 License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Model-as-a-service operators with over US$10 billion in 12-month revenue must pass a Z.ai security review before commercial use.","hugging_face":"https://huggingface.co/zai-org/GLM-5.3","article_fact_checked":true},{"name":"dots3-note Preview","article":"https://aiwiki.ai/wiki/dots3_note","kind":"language","developer":"dots studio (Xiaohongshu)","released":"2026-08-13","parameters":"280B total, 16B active (language backbone)","parameters_billions":280,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":false,"note":null,"hugging_face":"https://huggingface.co/dots-studio/dots3-note-prev","article_fact_checked":false},{"name":"Nemotron 3.5 Lightning","article":"https://aiwiki.ai/wiki/nemotron_3_5_lightning","kind":"language","developer":"NVIDIA","released":"2026-08-11","parameters":"30B total, 3B active (31.6B / 3.6B counting embeddings)","parameters_billions":30,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":null,"training_data_published":true,"note":"NVIDIA published pre-training and post-training dataset collections and recipes alongside the weights.","hugging_face":"https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16","article_fact_checked":true},{"name":"Muse Glimmer","article":"https://aiwiki.ai/wiki/muse_glimmer","kind":"language","developer":"Meta AI","released":"2026-08-10","parameters":"About 29.6B (29,776,626,688 in the BF16 checkpoint)","parameters_billions":29.6,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Meta also publishes a separate Muse Glimmer Usage Policy that lists prohibited uses.","hugging_face":"https://huggingface.co/meta-models/Muse-Glimmer-30B","article_fact_checked":true},{"name":"MAGI-2 Preview","article":"https://aiwiki.ai/wiki/magi_2","kind":"video","developer":"Sand.ai","released":"2026-08-05","parameters":"About 114B total, about 6B active","parameters_billions":114,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":false,"training_data_published":false,"note":"The bundled audio VAE is Stable Audio Open 1.0, which upstream carries the Stable Audio Community License.","hugging_face":"https://huggingface.co/sand-ai/MAGI-2-preview","article_fact_checked":false},{"name":"Qwen3.8-Max","article":"https://aiwiki.ai/wiki/qwen3_8_max","kind":"language","developer":"Qwen team, Alibaba","released":"2026-08-03","parameters":"2.4T total, 95B active","parameters_billions":2400,"tier":"conditional","weights":"open","license":"custom","license_name":"Qwen3.8-Max License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Open checkpoint is text only and requires thinking mode. Model-as-a-service firms above US$50M yearly revenue need a separate Qwen license.","hugging_face":"https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B","article_fact_checked":true},{"name":"BigBang-v1","article":"https://aiwiki.ai/wiki/bigbang_v1","kind":"language","developer":"The Endless Frontier team","released":"2026-08-02","parameters":"35B total, about 3B active","parameters_billions":35,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/endless-frontier/BigBang-v1","article_fact_checked":true},{"name":"MiniMax H3","article":"https://aiwiki.ai/wiki/minimax_h3","kind":"video","developer":"MiniMax","released":"2026-07-31","parameters":"33B (H3-Omni-Transformer generator)","parameters_billions":33,"tier":"conditional","weights":"open","license":"custom","license_name":"MiniMax H3 Community License Agreement","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"License excludes the EU, UK, South Korea and US. Over $20M yearly revenue needs authorization. Context-IR and Regenerate-2K stages are not released.","hugging_face":"https://huggingface.co/MiniMaxAI/MiniMax-H3","article_fact_checked":true},{"name":"Frontis-MA1","article":"https://aiwiki.ai/wiki/frontis_ma1","kind":"other","developer":"FrontisAI (Horizon Research of Frontis.AI with Tsinghua University)","released":"2026-07-30","parameters":"35B total, about 3B active (Frontis-MA1-35B)","parameters_billions":35,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":true,"note":null,"hugging_face":"https://huggingface.co/FrontisAI/Frontis-MA1-35B","article_fact_checked":true},{"name":"Ling-3.0-flash","article":"https://aiwiki.ai/wiki/ling_3_0_flash","kind":"language","developer":"inclusionAI","released":"2026-07-23","parameters":"124B total, 5.1B active per token","parameters_billions":124,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/inclusionAI/Ling-3.0-flash","article_fact_checked":true},{"name":"Qwen3.8","article":"https://aiwiki.ai/wiki/qwen3_8","kind":"language","developer":"Qwen team, Alibaba Cloud","released":"2026-07-19","parameters":"27.8B (Qwen3.8-27B); 2.4T total, 95B active (Qwen3.8-Max)","parameters_billions":27.8,"tier":"conditional","weights":"open","license":"custom","license_name":"Qwen3.8-Max License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Applies to the flagship Qwen3.8-2.4T-A95B; Qwen3.8-27B is Apache 2.0. Hosted qwen3.8-max snapshots add features the open checkpoint lacks.","hugging_face":null,"article_fact_checked":true},{"name":"Kimi K3","article":"https://aiwiki.ai/wiki/kimi_k3","kind":"language","developer":"Moonshot AI","released":"2026-07-16","parameters":"2.78T total; 104.2B activated per token","parameters_billions":2780,"tier":"conditional","weights":"open","license":"custom","license_name":"Kimi K3 License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Model-as-a-service firms over $20M yearly revenue need a separate agreement; very large products must display Kimi K3.","hugging_face":"https://huggingface.co/moonshotai/Kimi-K3","article_fact_checked":true},{"name":"Inkling (Thinking Machines Lab)","article":"https://aiwiki.ai/wiki/inkling","kind":"language","developer":"Thinking Machines Lab","released":"2026-07-15","parameters":"975B total, 41B active","parameters_billions":975,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":false,"training_data_published":false,"note":"A separate Model Acceptable Use Policy restricts some uses. Inkling-Small is also open under Apache 2.0.","hugging_face":"https://huggingface.co/thinkingmachines/Inkling","article_fact_checked":true},{"name":"LingBot-VLA 2.0","article":"https://aiwiki.ai/wiki/lingbot_vla_2_0","kind":"robotics","developer":"Robbyant (Ant Group)","released":"2026-07","parameters":"6B","parameters_billions":6,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":false,"note":"The Qwen3-VL backbone and other dependencies keep their own license terms.","hugging_face":"https://huggingface.co/robbyant/lingbot-vla-v2-6b","article_fact_checked":true},{"name":"Leanstral 1.5","article":"https://aiwiki.ai/wiki/leanstral_1_5","kind":"language","developer":"Mistral AI","released":"2026-06-30","parameters":"119B total, 6.5B active","parameters_billions":119,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":false,"training_data_published":false,"note":null,"hugging_face":"https://huggingface.co/mistralai/Leanstral-1.5-119B-A6B","article_fact_checked":false},{"name":"NVIDIA STRATA","article":"https://aiwiki.ai/wiki/nvidia_strata","kind":"science","developer":"NVIDIA, with U.S. national laboratories","released":"2026-06-30","parameters":"331.7M","parameters_billions":0.3317,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":true,"training_data_published":true,"note":null,"hugging_face":"https://huggingface.co/nvidia/strata","article_fact_checked":false},{"name":"Kimi K2.7-Code","article":"https://aiwiki.ai/wiki/kimi_k2_7_code","kind":"language","developer":"Moonshot AI","released":"2026-06-12","parameters":"1T total, 32B active","parameters_billions":1000,"tier":"conditional","weights":"open","license":"modified-mit","license_name":"Modified MIT","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"MIT plus one clause: products over 100M monthly active users or $20M monthly revenue must display Kimi K2.7 Code in the user interface.","hugging_face":"https://huggingface.co/moonshotai/Kimi-K2.7-Code","article_fact_checked":true},{"name":"DiffusionGemma","article":"https://aiwiki.ai/wiki/diffusiongemma","kind":"language","developer":"Google DeepMind","released":"2026-06-10","parameters":"25.2B total, 3.8B active","parameters_billions":25.2,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":false,"note":null,"hugging_face":"https://huggingface.co/google/diffusiongemma-26B-A4B-it","article_fact_checked":true},{"name":"North Mini Code","article":"https://aiwiki.ai/wiki/north_mini_code","kind":"language","developer":"Cohere","released":"2026-06-09","parameters":"30B total, about 3B active","parameters_billions":30,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":false,"training_data_published":false,"note":null,"hugging_face":"https://huggingface.co/CohereLabs/North-Mini-Code-1.0","article_fact_checked":false},{"name":"MiniMax M3","article":"https://aiwiki.ai/wiki/minimax_m3","kind":"language","developer":"MiniMax","released":"2026-06-01","parameters":"About 427B (427,040,140,160 in the published checkpoint)","parameters_billions":427,"tier":"conditional","weights":"open","license":"custom","license_name":"MiniMax Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Commercial use needs a Built with MiniMax M3 notice and a notice to MiniMax; written authorization above $20M yearly revenue.","hugging_face":"https://huggingface.co/MiniMaxAI/MiniMax-M3","article_fact_checked":true},{"name":"NVIDIA Alpamayo 2 Super","article":"https://aiwiki.ai/wiki/nvidia_alpamayo_2_super","kind":"robotics","developer":"NVIDIA","released":"2026-06-01","parameters":"34 billion (32B vision-language backbone plus a diffusion Action Expert of about 2B); announced in June 2026 as 32 billion","parameters_billions":34,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Inference code is Apache 2.0.","hugging_face":"https://huggingface.co/nvidia/Alpamayo2-Super","article_fact_checked":true},{"name":"NVIDIA Cosmos 3","article":"https://aiwiki.ai/wiki/nvidia_cosmos_3","kind":"robotics","developer":"NVIDIA","released":"2026-06-01","parameters":"4B (Edge), 16B (Nano), 64B (Super)","parameters_billions":4,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/nvidia/Cosmos3-Super","article_fact_checked":true},{"name":"GLM-5.2","article":"https://aiwiki.ai/wiki/glm_5_2","kind":"language","developer":"Zhipu AI (Z.ai)","released":"2026-06","parameters":"744B total, 40B active per Z.ai's model index (Hugging Face counts about 753B, including the multi-token prediction layer)","parameters_billions":744,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/zai-org/GLM-5.2","article_fact_checked":true},{"name":"Hy Memory","article":"https://aiwiki.ai/wiki/hy_memory","kind":"language","developer":"Tencent Hunyuan (Tencent Hy)","released":"2026-05-28","parameters":"About 3B active (Hy-Memory-30B-A3B)","parameters_billions":3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"The page covers a memory system; its Hy-Memory-30B-A3B model, the SDK and the MemHub app are all Apache 2.0.","hugging_face":"https://huggingface.co/tencent/Hy-Memory-30B-A3B","article_fact_checked":true},{"name":"ZAYA1-8B","article":"https://aiwiki.ai/wiki/zaya1_8b","kind":"language","developer":"Zyphra","released":"2026-05-06","parameters":"8.4B total; 760M active per token","parameters_billions":8.4,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Zyphra/ZAYA1-8B","article_fact_checked":false},{"name":"Mistral Medium 3.5","article":"https://aiwiki.ai/wiki/mistral_medium_3_5","kind":"language","developer":"Mistral AI","released":"2026-04-28","parameters":"~128 billion (all active per token)","parameters_billions":128,"tier":"conditional","weights":"open","license":"modified-mit","license_name":"Modified MIT","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Not licensed to companies with more than $20M monthly revenue, which need a commercial license from Mistral.","hugging_face":"https://huggingface.co/mistralai/Mistral-Medium-3.5-128B","article_fact_checked":false},{"name":"DeepSeek V4","article":"https://aiwiki.ai/wiki/deepseek_v4","kind":"language","developer":"DeepSeek","released":"2026-04-24","parameters":"1.6T total, 49B active (Pro); 284B total, 13B active (Flash)","parameters_billions":1600,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro","article_fact_checked":true},{"name":"DeepSeek V4-Flash","article":"https://aiwiki.ai/wiki/deepseek_v4_flash","kind":"language","developer":"DeepSeek","released":"2026-04-24","parameters":"284B total, 13B active","parameters_billions":284,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Preview, official 0731 and Vision-Exp checkpoints are all MIT; the hosted V4-Flash API names were retired on September 10, 2026.","hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731","article_fact_checked":true},{"name":"DeepSeek V4-Pro","article":"https://aiwiki.ai/wiki/deepseek_v4_pro","kind":"language","developer":"DeepSeek","released":"2026-04-24","parameters":"1.6T total, 49B active","parameters_billions":1600,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813","article_fact_checked":true},{"name":"Tencent Hunyuan Hy3","article":"https://aiwiki.ai/wiki/hunyuan_hy3","kind":"language","developer":"Tencent","released":"2026-04-23","parameters":"~295B total; ~21B active per token","parameters_billions":295,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Hy3-preview (April 2026) used the Tencent Hy Community License; the official Hy3 weights (July 2026) are Apache 2.0.","hugging_face":"https://huggingface.co/tencent/Hy3","article_fact_checked":true},{"name":"Kimi K2.6","article":"https://aiwiki.ai/wiki/kimi_k2_6","kind":"language","developer":"Moonshot AI","released":"2026-04-20","parameters":"~1T total; ~32B active per token","parameters_billions":1000,"tier":"conditional","weights":"open","license":"modified-mit","license_name":"Modified MIT","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"MIT plus one clause: products over 100M monthly active users or $20M monthly revenue must display Kimi K2.6 in the user interface.","hugging_face":"https://huggingface.co/moonshotai/Kimi-K2.6","article_fact_checked":true},{"name":"NVIDIA Isaac GR00T N1.7","article":"https://aiwiki.ai/wiki/groot_n1_7","kind":"robotics","developer":"NVIDIA","released":"2026-04-17","parameters":"3B","parameters_billions":3,"tier":"conditional","weights":"open","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Model card and bundled LICENSE give the NVIDIA Open Model License; two subcards still link an older noncommercial license. Code is Apache 2.0.","hugging_face":"https://huggingface.co/nvidia/GR00T-N1.7-3B","article_fact_checked":true},{"name":"HY-World 2.0","article":"https://aiwiki.ai/wiki/hy_world_2","kind":"vision","developer":"Tencent Hunyuan","released":"2026-04-16","parameters":"About 425M to 80B per component (WorldMirror 2.0 1.2B, WorldStereo 2.0 17B, HY-Pano 2.0 80B)","parameters_billions":0.425,"tier":"conditional","weights":"open","license":"hunyuan-community","license_name":"Tencent Hunyuan Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"The license does not apply in the EU, the UK and South Korea, and carries an acceptable use policy.","hugging_face":"https://huggingface.co/tencent/HY-World-2.0","article_fact_checked":false},{"name":"Lyra 2.0","article":"https://aiwiki.ai/wiki/lyra_2_0","kind":"vision","developer":"NVIDIA Spatial Intelligence Lab","released":"2026-04-15","parameters":null,"parameters_billions":null,"tier":"noncommercial","weights":"open","license":"custom","license_name":"NVIDIA Internal Scientific R&D Model License","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Internal research and development only: no distribution, deployment or production use. The code is Apache 2.0.","hugging_face":"https://huggingface.co/nvidia/Lyra-2.0","article_fact_checked":false},{"name":"NVIDIA Ising","article":"https://aiwiki.ai/wiki/nvidia_ising","kind":"other","developer":"NVIDIA","released":"2026-04-14","parameters":"31B (Calibration 1.5); 35B total, 3B active (Calibration 1)","parameters_billions":31,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Calibration 1.5 and the ColorCode decoder are OpenMDW 1.1; Calibration 1 and the SurfaceCode decoders use the NVIDIA Open Model License. Decoders are gated.","hugging_face":null,"article_fact_checked":true},{"name":"Muse Spark","article":"https://aiwiki.ai/wiki/muse_spark","kind":"language","developer":"Meta Superintelligence Labs","released":"2026-04-08","parameters":"816B total, 42B active (Muse Spark Open)","parameters_billions":816,"tier":"partial","weights":"partial","license":"custom","license_name":"Muse Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Muse Spark 1.0 to 1.3 are hosted only. Muse Spark Open (October 2026) has gated weights under the Muse Model License.","hugging_face":null,"article_fact_checked":true},{"name":"GLM-5.1","article":"https://aiwiki.ai/wiki/glm_5_1","kind":"language","developer":"Zhipu AI (Z.ai)","released":"2026-04-07","parameters":"744B total, 40B active per Z.ai's model index (Hugging Face counts about 754B, including the multi-token prediction layer)","parameters_billions":744,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/zai-org/GLM-5.1","article_fact_checked":true},{"name":"Gemma 4","article":"https://aiwiki.ai/wiki/gemma_4","kind":"language","developer":"Google DeepMind","released":"2026-04-02","parameters":"E2B 5.1B; E4B 8B; 26B A4B 25.2B (3.8B active); 31B 30.7B","parameters_billions":2,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/google/gemma-4-31B-it","article_fact_checked":true},{"name":"Hy-Embodied","article":"https://aiwiki.ai/wiki/hy_embodied","kind":"robotics","developer":"Tencent (Hunyuan team and Tencent Robotics X)","released":"2026-04","parameters":"About 4B (VLA-0.5), 6.2B (RxBrain-1.0), 30B total with about 3B activated (VLM-1.0)","parameters_billions":4,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"VLM-1.0, RxBrain-1.0 and VLA-0.5 are Apache 2.0. The earlier HY-Embodied-0.5 and 0.5-X use the Tencent HY Community License (excludes EU, UK, South Korea).","hugging_face":null,"article_fact_checked":false},{"name":"Qwen3.6","article":"https://aiwiki.ai/wiki/qwen3_6","kind":"language","developer":"Alibaba (Qwen team)","released":"2026-04","parameters":"Open: 35B/3B active (MoE), 27B (dense); hosted undisclosed","parameters_billions":35,"tier":"partial","weights":"partial","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Qwen3.6-35B-A3B and Qwen3.6-27B are Apache 2.0; Qwen3.6-Plus, Flash and Max-Preview are hosted only.","hugging_face":"https://huggingface.co/Qwen/Qwen3.6-35B-A3B","article_fact_checked":true},{"name":"Xiaomi MiMo-V2.5","article":"https://aiwiki.ai/wiki/mimo_v2_5","kind":"other","developer":"Xiaomi","released":"2026-04","parameters":"310B total, 15B active (standard); 1.02T total, 42B active (Pro)","parameters_billions":310,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":true},{"name":"Cohere Transcribe","article":"https://aiwiki.ai/wiki/cohere_transcribe","kind":"audio","developer":"Cohere and Cohere Labs","released":"2026-03-26","parameters":"About 2.07B (2,065,804,048 in the published checkpoint)","parameters_billions":2.07,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/CohereLabs/cohere-transcribe-03-2026","article_fact_checked":true},{"name":"MiniMax M2.7","article":"https://aiwiki.ai/wiki/minimax_m2_7","kind":"language","developer":"MiniMax","released":"2026-03-18","parameters":"~230B total; ~10B active per token","parameters_billions":230,"tier":"noncommercial","weights":"open","license":"custom","license_name":"MiniMax M2.7 Non-Commercial License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"MIT-style text titled Non-Commercial License. Commercial use needs prior written authorization from MiniMax.","hugging_face":"https://huggingface.co/MiniMaxAI/MiniMax-M2.7","article_fact_checked":false},{"name":"Mistral Small 4","article":"https://aiwiki.ai/wiki/mistral_small_4","kind":"language","developer":"Mistral AI","released":"2026-03-16","parameters":"119 billion total; ~6.5 billion active per token","parameters_billions":119,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/mistralai/Mistral-Small-4-119B-2603","article_fact_checked":false},{"name":"Qwen3.5","article":"https://aiwiki.ai/wiki/qwen3_5","kind":"language","developer":"Qwen team, Alibaba","released":"2026-02-16","parameters":"397B total; about 17B active per token (flagship Qwen3.5-397B-A17B)","parameters_billions":397,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/Qwen3.5-397B-A17B","article_fact_checked":true},{"name":"GLM-5","article":"https://aiwiki.ai/wiki/glm_5","kind":"language","developer":"Zhipu AI (Z.ai)","released":"2026-02-11","parameters":"744B total; ~40B active per token","parameters_billions":744,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/zai-org/GLM-5","article_fact_checked":true},{"name":"Qwen3-Coder-Next","article":"https://aiwiki.ai/wiki/qwen3_coder_next","kind":"language","developer":"Alibaba (Qwen team)","released":"2026-02-03","parameters":"About 80B total, about 3B active per token","parameters_billions":80,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/Qwen3-Coder-Next","article_fact_checked":false},{"name":"Kimi K2.5","article":"https://aiwiki.ai/wiki/kimi_k2_5","kind":"language","developer":"Moonshot AI","released":"2026-01-27","parameters":"1T total; 32B active per token","parameters_billions":1000,"tier":"conditional","weights":"open","license":"modified-mit","license_name":"Modified MIT","commercial_use":"conditional","training_code_published":null,"training_data_published":false,"note":"MIT plus one clause: products over 100M monthly active users or $20M monthly revenue must display Kimi K2.5 in the user interface.","hugging_face":"https://huggingface.co/moonshotai/Kimi-K2.5","article_fact_checked":true},{"name":"Jamba2","article":"https://aiwiki.ai/wiki/jamba_2","kind":"language","developer":"AI21 Labs","released":"2026-01-08","parameters":"3B dense; Mini 52B total, 12B active","parameters_billions":3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/ai21labs/AI21-Jamba2-Mini","article_fact_checked":false},{"name":"Nemotron 3","article":"https://aiwiki.ai/wiki/nemotron_3","kind":"language","developer":"NVIDIA","released":"2025-12-15","parameters":"Nano 30B (3.5B active); Super 120.6B (12.7B); Ultra 550B (55B)","parameters_billions":30,"tier":"conditional","weights":"open","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Nano and Super use the NVIDIA Nemotron Open Model License. Ultra (June 2026) is under the permissive OpenMDW 1.1.","hugging_face":null,"article_fact_checked":true},{"name":"NVIDIA Isaac GR00T N1.6","article":"https://aiwiki.ai/wiki/groot_n1_6","kind":"robotics","developer":"NVIDIA (GEAR team)","released":"2025-12-15","parameters":"3B","parameters_billions":3,"tier":"noncommercial","weights":"open","license":"custom","license_name":"NVIDIA License (non-commercial)","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Weights are limited to noncommercial research or evaluation (except NVIDIA and affiliates); the code is Apache 2.0.","hugging_face":"https://huggingface.co/nvidia/GR00T-N1.6-3B","article_fact_checked":false},{"name":"Mistral Large 3","article":"https://aiwiki.ai/wiki/mistral_large_3","kind":"language","developer":"Mistral AI","released":"2025-12-02","parameters":"~675B total; ~41B active per token (vision encoder ~2.5B)","parameters_billions":675,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/mistralai/Mistral-Large-3-675B-Instruct-2512","article_fact_checked":true},{"name":"FLUX.2","article":"https://aiwiki.ai/wiki/flux_2","kind":"image","developer":"Black Forest Labs","released":"2025-11-25","parameters":"~32B (dev, pro, flex, max); klein 4B and 9B","parameters_billions":32,"tier":"partial","weights":"partial","license":"flux-non-commercial","license_name":"FLUX Non-Commercial License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"[dev] and klein 9B are non-commercial; klein 4B is Apache 2.0; [pro], [flex] and [max] are API only. BFL sells commercial licenses.","hugging_face":"https://huggingface.co/black-forest-labs/FLUX.2-dev","article_fact_checked":true},{"name":"OLMo 3","article":"https://aiwiki.ai/wiki/olmo_3","kind":"language","developer":"Allen Institute for AI (Ai2)","released":"2025-11-20","parameters":"7B and 32B","parameters_billions":7,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/allenai/Olmo-3-1125-32B","article_fact_checked":false},{"name":"Omnilingual ASR","article":"https://aiwiki.ai/wiki/omnilingual_asr","kind":"audio","developer":"Meta FAIR","released":"2025-11-10","parameters":"0.32B to 7.81B","parameters_billions":0.32,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"The commissioned Omnilingual ASR Corpus is published under CC BY 4.0; the rest of the training data is not.","hugging_face":null,"article_fact_checked":true},{"name":"Kimi K2 Thinking","article":"https://aiwiki.ai/wiki/kimi_k2_thinking","kind":"language","developer":"Moonshot AI","released":"2025-11-06","parameters":"About 1 trillion total; about 32 billion active per token","parameters_billions":1000,"tier":"conditional","weights":"open","license":"modified-mit","license_name":"Modified MIT","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"MIT plus an attribution clause for products over 100M monthly active users or $20M monthly revenue.","hugging_face":"https://huggingface.co/moonshotai/Kimi-K2-Thinking","article_fact_checked":true},{"name":"Kimi Linear","article":"https://aiwiki.ai/wiki/kimi_linear","kind":"language","developer":"Moonshot AI","released":"2025-10-30","parameters":"48B total, 3B active","parameters_billions":48,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/moonshotai/Kimi-Linear-48B-A3B-Instruct","article_fact_checked":false},{"name":"MiniMax M2","article":"https://aiwiki.ai/wiki/minimax_m2","kind":"language","developer":"MiniMax","released":"2025-10-27","parameters":"230B total; ~10B active per token","parameters_billions":230,"tier":"conditional","weights":"open","license":"modified-mit","license_name":"Modified MIT","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"MIT at launch, amended Nov 13, 2025: products over 100M monthly users or $30M annual revenue must display MiniMax M2 in the interface.","hugging_face":"https://huggingface.co/MiniMaxAI/MiniMax-M2","article_fact_checked":true},{"name":"DeepSeek-OCR","article":"https://aiwiki.ai/wiki/deepseek-ocr","kind":"vision","developer":"DeepSeek","released":"2025-10-20","parameters":"About 3B decoder (about 570M activated per token) plus about 380M encoder","parameters_billions":3,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-OCR","article_fact_checked":false},{"name":"Ling-1T","article":"https://aiwiki.ai/wiki/ling_1t","kind":"language","developer":"Ant Group (Inclusion AI)","released":"2025-10-09","parameters":"~1 trillion total; 51B active per token","parameters_billions":1000,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/inclusionAI/Ling-1T","article_fact_checked":true},{"name":"Jamba Reasoning 3B","article":"https://aiwiki.ai/wiki/jamba_reasoning_3b","kind":"language","developer":"AI21 Labs","released":"2025-10-08","parameters":"~3B","parameters_billions":3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/ai21labs/AI21-Jamba-Reasoning-3B","article_fact_checked":false},{"name":"IBM Granite 4.0","article":"https://aiwiki.ai/wiki/granite_4","kind":"language","developer":"IBM","released":"2025-10-02","parameters":"~3B (Micro) to ~32B total with ~9B active (H-Small)","parameters_billions":3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/ibm-granite/granite-4.0-h-small","article_fact_checked":false},{"name":"Ring-1T","article":"https://aiwiki.ai/wiki/ring_1t","kind":"language","developer":"inclusionAI (Ant Group)","released":"2025-10","parameters":"~1T total; ~50B active per token","parameters_billions":1000,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/inclusionAI/Ring-1T","article_fact_checked":false},{"name":"GLM-4.6","article":"https://aiwiki.ai/wiki/glm_4_6","kind":"language","developer":"Zhipu AI (Z.ai)","released":"2025-09-30","parameters":"~357B total; ~32B active per token","parameters_billions":357,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/zai-org/GLM-4.6","article_fact_checked":true},{"name":"DeepSeek V3.2","article":"https://aiwiki.ai/wiki/deepseek_v3_2","kind":"language","developer":"DeepSeek","released":"2025-09-29","parameters":"671B total; 37B active per token","parameters_billions":671,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-V3.2","article_fact_checked":true},{"name":"Hunyuan Image 3.0","article":"https://aiwiki.ai/wiki/hunyuan_image_3","kind":"image","developer":"Tencent","released":"2025-09-28","parameters":"~80B total; ~13B active per token; 64 experts","parameters_billions":80,"tier":"conditional","weights":"open","license":"hunyuan-community","license_name":"Tencent Hunyuan Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Excludes the EU, UK and South Korea; above 100 million monthly active users a separate license is required.","hugging_face":"https://huggingface.co/tencent/HunyuanImage-3.0","article_fact_checked":false},{"name":"Qwen3-VL","article":"https://aiwiki.ai/wiki/qwen3_vl","kind":"language","developer":"Qwen team, Alibaba Cloud","released":"2025-09-23","parameters":"Dense 2B, 4B, 8B, 32B; MoE 30B-A3B and 235B-A22B","parameters_billions":2,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/Qwen3-VL-235B-A22B-Instruct","article_fact_checked":true},{"name":"EmbeddingGemma","article":"https://aiwiki.ai/wiki/embeddinggemma","kind":"language","developer":"Google DeepMind","released":"2025-09-04","parameters":"~308M (~100M backbone, ~200M embedding table)","parameters_billions":0.308,"tier":"conditional","weights":"open","license":"gemma-terms","license_name":"Gemma Terms of Use","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated on Hugging Face: accept the Gemma Terms of Use first. Commercial use allowed subject to the prohibited use policy.","hugging_face":"https://huggingface.co/google/embeddinggemma-300m","article_fact_checked":false},{"name":"Apertus","article":"https://aiwiki.ai/wiki/apertus","kind":"language","developer":"Swiss AI Initiative (EPFL, ETH Zurich, CSCS)","released":"2025-09-02","parameters":"8B and 70B (70B reported as 71B); base and Instruct variants","parameters_billions":8,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/swiss-ai/Apertus-70B-2509","article_fact_checked":false},{"name":"LongCat-Flash","article":"https://aiwiki.ai/wiki/longcat_flash","kind":"language","developer":"Meituan (LongCat team)","released":"2025-09-01","parameters":"~560B total; 18.6B-31.3B active per token (avg ~27B)","parameters_billions":560,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/meituan-longcat/LongCat-Flash-Chat","article_fact_checked":false},{"name":"Qwen3-Next","article":"https://aiwiki.ai/wiki/qwen3_next","kind":"language","developer":"Qwen team, Alibaba Cloud","released":"2025-09","parameters":"80 billion total; ~3 billion active per token","parameters_billions":80,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/Qwen3-Next-80B-A3B-Instruct","article_fact_checked":false},{"name":"Qwen3-Omni","article":"https://aiwiki.ai/wiki/qwen3_omni","kind":"language","developer":"Qwen team, Alibaba Cloud","released":"2025-09","parameters":"30B total (Thinker); ~3B active per token","parameters_billions":30,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/Qwen3-Omni-30B-A3B-Instruct","article_fact_checked":false},{"name":"Command A Reasoning","article":"https://aiwiki.ai/wiki/command_a_reasoning","kind":"language","developer":"Cohere and Cohere Labs","released":"2025-08-21","parameters":"111 billion","parameters_billions":111,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Gated on Hugging Face (contact details required). Commercial use only through Cohere.","hugging_face":"https://huggingface.co/CohereLabs/command-a-reasoning-08-2025","article_fact_checked":false},{"name":"Jet-Nemotron","article":"https://aiwiki.ai/wiki/jet_nemotron","kind":"language","developer":"NVIDIA Research","released":"2025-08-21","parameters":"~2B and ~4B variants","parameters_billions":2,"tier":"noncommercial","weights":"open","license":"custom","license_name":"NVIDIA License (non-commercial)","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Weights may be used for research or evaluation only; the code is Apache 2.0.","hugging_face":"https://huggingface.co/jet-ai/Jet-Nemotron-2B","article_fact_checked":false},{"name":"rBio (CZI)","article":"https://aiwiki.ai/wiki/czi_rbio","kind":"science","developer":"Chan Zuckerberg Initiative (CZI)","released":"2025-08-21","parameters":"3 billion (base model)","parameters_billions":3,"tier":"noncommercial","weights":"open","license":"qwen-research","license_name":"Qwen Research License","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Built on Qwen2.5-3B-Instruct, so the Qwen Research License applies; CZI's own changes are MIT. Weights are on AWS S3, not Hugging Face.","hugging_face":null,"article_fact_checked":false},{"name":"DeepSeek V3.1","article":"https://aiwiki.ai/wiki/deepseek_v3_1","kind":"language","developer":"DeepSeek","released":"2025-08-19","parameters":"671B total, 37B active per token (685B checkpoint incl. ~14B MTP module)","parameters_billions":671,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-V3.1","article_fact_checked":false},{"name":"Nemotron Nano 2","article":"https://aiwiki.ai/wiki/nemotron_nano_2","kind":"language","developer":"NVIDIA","released":"2025-08-18","parameters":"~9B (flagship, pruned and distilled from a 12B base)","parameters_billions":9,"tier":"conditional","weights":"open","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":true,"note":"Most of the pre-training corpus is published as Nemotron-Pre-Training-Dataset-v1 under CC BY 4.0.","hugging_face":"https://huggingface.co/nvidia/NVIDIA-Nemotron-Nano-9B-v2","article_fact_checked":false},{"name":"DINOv3","article":"https://aiwiki.ai/wiki/dinov3","kind":"vision","developer":"Meta AI (FAIR)","released":"2025-08-14","parameters":"21M to 6.7B (ViT-7B/16 largest)","parameters_billions":0.021,"tier":"conditional","weights":"open","license":"custom","license_name":"DINOv3 License","commercial_use":"conditional","training_code_published":true,"training_data_published":null,"note":"Gated on Hugging Face (manual approval). Commercial use allowed; the license bars ITAR-controlled and trade-control-prohibited end uses.","hugging_face":"https://huggingface.co/facebook/dinov3-vit7b16-pretrain-lvd1689m","article_fact_checked":false},{"name":"gpt-oss","article":"https://aiwiki.ai/wiki/gpt_oss","kind":"language","developer":"OpenAI","released":"2025-08-05","parameters":"120b: 117B total, 5.1B active; 20b: 21B total, 3.6B active","parameters_billions":120,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/openai/gpt-oss-120b","article_fact_checked":true},{"name":"Qwen-Image","article":"https://aiwiki.ai/wiki/qwen_image","kind":"other","developer":"Qwen team, Alibaba (Tongyi Lab)","released":"2025-08-04","parameters":"~20 billion (MMDiT backbone)","parameters_billions":20,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/Qwen-Image","article_fact_checked":true},{"name":"Hermes 4","article":"https://aiwiki.ai/wiki/hermes_4","kind":"language","developer":"Nous Research","released":"2025-08","parameters":"14B, 70B, 405B","parameters_billions":14,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"70B and 405B (Llama 3.1 base) use the Llama 3.1 Community License; the 14B (Qwen3 base) is Apache 2.0.","hugging_face":"https://huggingface.co/NousResearch/Hermes-4-405B","article_fact_checked":false},{"name":"FoxBrain","article":"https://aiwiki.ai/wiki/foxbrain","kind":"language","developer":"Hon Hai Research Institute (Foxconn)","released":"2025-07-31","parameters":"70B (v1.x)","parameters_billions":70,"tier":"partial","weights":"partial","license":"custom","license_name":"FoxBrain Model Usage and License Agreement","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"v1.2 (70B) is manually gated for academic and research organizations; the v1.0 preview is partner-only and no FoxBrain 2.0 weights are published.","hugging_face":"https://huggingface.co/FoxconnAI/Llama_3.1-FoxBrain-70B-V1.2","article_fact_checked":true},{"name":"Step-3","article":"https://aiwiki.ai/wiki/step_3","kind":"language","developer":"StepFun","released":"2025-07-31","parameters":"~321B total; ~38B active per token","parameters_billions":321,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/stepfun-ai/step3","article_fact_checked":true},{"name":"GLM-4.5","article":"https://aiwiki.ai/wiki/glm_4_5","kind":"language","developer":"Zhipu AI (Z.ai)","released":"2025-07-28","parameters":"355B total, 32B active; Air: 106B total, 12B active","parameters_billions":355,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/zai-org/GLM-4.5","article_fact_checked":true},{"name":"HunyuanWorld 1.0","article":"https://aiwiki.ai/wiki/hunyuanworld","kind":"vision","developer":"Tencent (Hunyuan team)","released":"2025-07-26","parameters":null,"parameters_billions":null,"tier":"conditional","weights":"open","license":"hunyuan-community","license_name":"Tencent Hunyuan Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"License excludes the EU, UK and South Korea; licensees above 1 million monthly active users must request a separate license from Tencent.","hugging_face":"https://huggingface.co/tencent/HunyuanWorld-1","article_fact_checked":false},{"name":"Qwen3-Coder","article":"https://aiwiki.ai/wiki/qwen_3_coder","kind":"language","developer":"Alibaba Qwen team (Tongyi Lab)","released":"2025-07-22","parameters":"480B total; 35B active per token (flagship)","parameters_billions":480,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/Qwen3-Coder-480B-A35B-Instruct","article_fact_checked":true},{"name":"Voxtral","article":"https://aiwiki.ai/wiki/voxtral","kind":"audio","developer":"Mistral AI","released":"2025-07-15","parameters":"Voxtral Small 24.3B; Voxtral Mini 4.7B","parameters_billions":24.3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/mistralai/Voxtral-Small-24B-2507","article_fact_checked":false},{"name":"Kimi K2","article":"https://aiwiki.ai/wiki/kimi_k2","kind":"language","developer":"Moonshot AI","released":"2025-07-11","parameters":"1.04T total; 32.6B active per token","parameters_billions":1040,"tier":"conditional","weights":"open","license":"modified-mit","license_name":"Modified MIT","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"MIT plus one clause: products over 100M monthly active users or $20M monthly revenue must display Kimi K2 in the user interface.","hugging_face":"https://huggingface.co/moonshotai/Kimi-K2-Instruct","article_fact_checked":true},{"name":"SmolLM 3","article":"https://aiwiki.ai/wiki/smollm_3","kind":"language","developer":"Hugging Face","released":"2025-07-08","parameters":"3B","parameters_billions":3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/HuggingFaceTB/SmolLM3-3B","article_fact_checked":false},{"name":"Phi-4-mini-flash-reasoning","article":"https://aiwiki.ai/wiki/phi_4_mini_flash_reasoning","kind":"language","developer":"Microsoft","released":"2025-07","parameters":"3.8 billion","parameters_billions":3.8,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/microsoft/Phi-4-mini-flash-reasoning","article_fact_checked":false},{"name":"UMA (Universal Model for Atoms)","article":"https://aiwiki.ai/wiki/uma_meta","kind":"science","developer":"FAIR Chemistry team, Meta AI","released":"2025-06-30","parameters":"UMA-S ~150M, UMA-M ~1.4B, UMA-L ~700M (total)","parameters_billions":0.15,"tier":"conditional","weights":"open","license":"custom","license_name":"FAIR Chemistry License v1","commercial_use":"conditional","training_code_published":null,"training_data_published":true,"note":"Gated on Hugging Face with manual approval; commercial use allowed under an acceptable use policy that excludes some jurisdictions.","hugging_face":"https://huggingface.co/facebook/UMA","article_fact_checked":false},{"name":"AlphaGenome","article":"https://aiwiki.ai/wiki/alphagenome","kind":"science","developer":"Google DeepMind","released":"2025-06-25","parameters":null,"parameters_billions":null,"tier":"noncommercial","weights":"open","license":"custom","license_name":"AlphaGenome Model Terms","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Weights on Kaggle and Hugging Face after accepting non-commercial model terms; the code is Apache 2.0. Commercial use goes through Google Cloud.","hugging_face":null,"article_fact_checked":false},{"name":"State (Arc Institute virtual cell model)","article":"https://aiwiki.ai/wiki/arc_state_model","kind":"science","developer":"Arc Institute","released":"2025-06-23","parameters":"SE-600M checkpoint about 0.7B","parameters_billions":0.6,"tier":"noncommercial","weights":"open","license":"custom","license_name":"Arc Research Institute State Model Non-Commercial License","commercial_use":"no","training_code_published":true,"training_data_published":true,"note":"Weights are for non-commercial entities and purposes only; the code is CC BY-NC-SA 4.0.","hugging_face":null,"article_fact_checked":false},{"name":"MiniMax M1","article":"https://aiwiki.ai/wiki/minimax_m1","kind":"language","developer":"MiniMax","released":"2025-06-16","parameters":"456B total; 45.9B active per token","parameters_billions":456,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/MiniMaxAI/MiniMax-M1-80k","article_fact_checked":false},{"name":"V-JEPA 2","article":"https://aiwiki.ai/wiki/v_jepa_2","kind":"robotics","developer":"Meta AI","released":"2025-06-11","parameters":"1.2B (world-model configuration; encoder over 1B)","parameters_billions":1.2,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":true},{"name":"Magistral","article":"https://aiwiki.ai/wiki/magistral","kind":"language","developer":"Mistral AI","released":"2025-06-10","parameters":"24B (Magistral Small); unpublished (Magistral Medium)","parameters_billions":24,"tier":"partial","weights":"partial","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Magistral Small (24B) is open under Apache 2.0. Magistral Medium is available only through Mistral's API and products.","hugging_face":"https://huggingface.co/mistralai/Magistral-Small-2506","article_fact_checked":false},{"name":"SmolVLA","article":"https://aiwiki.ai/wiki/smolvla","kind":"robotics","developer":"Hugging Face (LeRobot team)","released":"2025-06-03","parameters":"~450M total (~100M action expert)","parameters_billions":0.45,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/lerobot/smolvla_base","article_fact_checked":true},{"name":"Boltz-2","article":"https://aiwiki.ai/wiki/boltz_2","kind":"science","developer":"MIT (Jameel Clinic, CSAIL) and Recursion","released":"2025-06","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":false},{"name":"Hunyuan-A13B","article":"https://aiwiki.ai/wiki/hunyuan_a13b","kind":"language","developer":"Tencent (Hunyuan team)","released":"2025-06","parameters":"~80B total; ~13B active per token","parameters_billions":80,"tier":"conditional","weights":"open","license":"hunyuan-community","license_name":"Tencent Hunyuan Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Excludes the EU, UK and South Korea; above 100 million monthly active users a separate license is required.","hugging_face":"https://huggingface.co/tencent/Hunyuan-A13B-Instruct","article_fact_checked":false},{"name":"Qwen3 Embedding","article":"https://aiwiki.ai/wiki/qwen3_embedding","kind":"language","developer":"Alibaba Qwen team","released":"2025-06","parameters":"0.6B, 4B, 8B","parameters_billions":0.6,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/Qwen3-Embedding-8B","article_fact_checked":false},{"name":"Devstral","article":"https://aiwiki.ai/wiki/devstral","kind":"language","developer":"Mistral AI with All Hands AI","released":"2025-05-21","parameters":"24 billion (Small); Medium not disclosed","parameters_billions":24,"tier":"partial","weights":"partial","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Devstral Small (2505, 2507, Small 2) is Apache 2.0; Devstral Medium is API only; Devstral 2 123B uses a modified MIT with a revenue cap.","hugging_face":"https://huggingface.co/mistralai/Devstral-Small-2505","article_fact_checked":false},{"name":"Falcon-H1","article":"https://aiwiki.ai/wiki/falcon_h1","kind":"language","developer":"Technology Innovation Institute (TII), Abu Dhabi","released":"2025-05-20","parameters":"Six sizes: 0.5B, 1.5B, 1.5B-Deep, 3B, 7B, 34B","parameters_billions":0.5,"tier":"conditional","weights":"open","license":"falcon-llm","license_name":"TII Falcon License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Apache-2.0-based TII Falcon LLM License with an acceptable use policy.","hugging_face":"https://huggingface.co/tiiuae/Falcon-H1-34B-Instruct","article_fact_checked":false},{"name":"Gemma 3n","article":"https://aiwiki.ai/wiki/gemma_3n","kind":"language","developer":"Google DeepMind","released":"2025-05-20","parameters":"5B (E2B) and 8B (E4B) raw; about 2B and 4B effective","parameters_billions":5,"tier":"conditional","weights":"open","license":"gemma-terms","license_name":"Gemma Terms of Use","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Hugging Face repositories are gated behind acceptance of the Gemma terms.","hugging_face":"https://huggingface.co/google/gemma-3n-E4B-it","article_fact_checked":false},{"name":"MedGemma","article":"https://aiwiki.ai/wiki/medgemma","kind":"science","developer":"Google","released":"2025-05-20","parameters":"4B and 27B","parameters_billions":4,"tier":"conditional","weights":"open","license":"custom","license_name":"Health AI Developer Foundations Terms of Use","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated: the HAI-DEF terms must be accepted. Use is bound by a prohibited use policy and health regulatory requirements.","hugging_face":null,"article_fact_checked":false},{"name":"Wan 2.1-VACE","article":"https://aiwiki.ai/wiki/wan_vace","kind":"video","developer":"Alibaba (Tongyi Lab)","released":"2025-05-14","parameters":"1.3B and 14B (two variants)","parameters_billions":1.3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Wan-AI/Wan2.1-VACE-14B","article_fact_checked":false},{"name":"Isaac GR00T N1.5","article":"https://aiwiki.ai/wiki/groot_n1_5","kind":"robotics","developer":"NVIDIA","released":"2025-05","parameters":"~3B (GR00T-N1.5-3B)","parameters_billions":3,"tier":"noncommercial","weights":"open","license":"custom","license_name":"NVIDIA License (non-commercial)","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Research or evaluation use only. N1.7 weights use the NVIDIA Open Model License; repository code is Apache 2.0.","hugging_face":"https://huggingface.co/nvidia/GR00T-N1.5-3B","article_fact_checked":false},{"name":"NVIDIA Cosmos Reason","article":"https://aiwiki.ai/wiki/cosmos_reason","kind":"robotics","developer":"NVIDIA","released":"2025-05","parameters":"7B (Cosmos-Reason1); 2B to 32B (Cosmos-Reason2)","parameters_billions":7,"tier":"conditional","weights":"open","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Cosmos-Reason1-7B and the Cosmos-Reason2 models are downloadable; the 56B Reason1 model was described in the paper only.","hugging_face":"https://huggingface.co/nvidia/Cosmos-Reason1-7B","article_fact_checked":false},{"name":"Phi-4 Reasoning","article":"https://aiwiki.ai/wiki/phi_4_reasoning","kind":"language","developer":"Microsoft","released":"2025-04-30","parameters":"14 billion","parameters_billions":14,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/microsoft/Phi-4-reasoning","article_fact_checked":false},{"name":"Qwen3","article":"https://aiwiki.ai/wiki/qwen_3","kind":"language","developer":"Qwen Team, Alibaba Cloud","released":"2025-04-29","parameters":"0.6B to 235B (flagship 235B total; 22B active)","parameters_billions":0.6,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/Qwen3-235B-A22B","article_fact_checked":true},{"name":"π0.5","article":"https://aiwiki.ai/wiki/pi_0_5","kind":"robotics","developer":"Physical Intelligence","released":"2025-04-22","parameters":"~3.3B (~3B PaliGemma backbone + ~300M action expert)","parameters_billions":3.3,"tier":"conditional","weights":"open","license":"gemma-terms","license_name":"Gemma Terms of Use","commercial_use":"conditional","training_code_published":true,"training_data_published":null,"note":"Released in openpi in September 2025 (base, LIBERO and DROID checkpoints). openpi is Apache 2.0 and ships the Gemma Terms for the PaliGemma-derived weights.","hugging_face":null,"article_fact_checked":true},{"name":"InternVL3","article":"https://aiwiki.ai/wiki/internvl3","kind":"language","developer":"OpenGVLab (Shanghai AI Laboratory)","released":"2025-04-11","parameters":"Seven dense sizes, 938M to 78.41B","parameters_billions":0.9380000000000001,"tier":"conditional","weights":"open","license":"qwen-license","license_name":"Qwen License (Tongyi Qianwen)","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Applies to the flagship InternVL3-78B (Qwen2.5-72B base); the 1B to 38B sizes are tagged Apache 2.0 and the 9B is MIT.","hugging_face":"https://huggingface.co/OpenGVLab/InternVL3-78B","article_fact_checked":false},{"name":"Llama 4","article":"https://aiwiki.ai/wiki/llama_4","kind":"language","developer":"Meta","released":"2025-04-05","parameters":"Scout: 109B total, 17B active; Maverick: ~400B total, 17B active","parameters_billions":109,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Scout and Maverick are downloadable (gated); Behemoth was never released. The license grant excludes EU-based individuals and companies.","hugging_face":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct","article_fact_checked":true},{"name":"Llama 4 Scout and Maverick","article":"https://aiwiki.ai/wiki/llama_4_scout_maverick","kind":"language","developer":"Meta","released":"2025-04-05","parameters":"Scout: ~109B total, 17B active (16 experts); Maverick: ~400B total, 17B active (128 experts)","parameters_billions":109,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated download. License grant excludes individuals and companies based in the EU; 700M monthly active user clause.","hugging_face":null,"article_fact_checked":false},{"name":"Apriel (ServiceNow)","article":"https://aiwiki.ai/wiki/apriel","kind":"language","developer":"ServiceNow (AI Research / SLAM lab)","released":"2025-04","parameters":"~4.8B to ~15B","parameters_billions":4.8,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/ServiceNow-AI/Apriel-1.5-15b-Thinker","article_fact_checked":false},{"name":"HiDream","article":"https://aiwiki.ai/wiki/hidream","kind":"image","developer":"HiDream.ai","released":"2025-04","parameters":"About 17 billion","parameters_billions":17,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/HiDream-ai/HiDream-I1-Full","article_fact_checked":false},{"name":"Nemotron-H","article":"https://aiwiki.ai/wiki/nemotron_h","kind":"language","developer":"NVIDIA","released":"2025-04","parameters":"8B, 47B, and 56B (base models)","parameters_billions":8,"tier":"noncommercial","weights":"open","license":"custom","license_name":"NVIDIA Internal Scientific R&D Model License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Research and development use only under NVIDIA's research license.","hugging_face":"https://huggingface.co/nvidia/Nemotron-H-56B-Base-8K","article_fact_checked":false},{"name":"TxGemma","article":"https://aiwiki.ai/wiki/txgemma","kind":"science","developer":"Google","released":"2025-03-25","parameters":"2B, 9B, 27B","parameters_billions":2,"tier":"conditional","weights":"open","license":"custom","license_name":"Health AI Developer Foundations Terms of Use","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated: the HAI-DEF terms must be accepted. Use is bound by a prohibited use policy and health regulatory requirements.","hugging_face":null,"article_fact_checked":false},{"name":"Llama Nemotron","article":"https://aiwiki.ai/wiki/llama_nemotron","kind":"language","developer":"Nvidia","released":"2025-03-18","parameters":"8B (Nano); 49B (Super); 253B (Ultra)","parameters_billions":8,"tier":"conditional","weights":"open","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":true,"note":"Also subject to the Llama 3.1 or 3.3 Community License. The post-training dataset is published; the Llama base data is not.","hugging_face":null,"article_fact_checked":false},{"name":"NVIDIA Isaac GR00T N1","article":"https://aiwiki.ai/wiki/groot_n1","kind":"robotics","developer":"NVIDIA (GEAR research lab)","released":"2025-03-18","parameters":"~2B (base N1)","parameters_billions":2,"tier":"noncommercial","weights":"open","license":"custom","license_name":"NVIDIA License (non-commercial)","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"N1, N1.5 and N1.6 weights are for research or evaluation only; N1.7 uses the NVIDIA Open Model License. Code is Apache 2.0.","hugging_face":"https://huggingface.co/nvidia/GR00T-N1-2B","article_fact_checked":false},{"name":"RWKV-7 (Goose)","article":"https://aiwiki.ai/wiki/rwkv_7","kind":"language","developer":"RWKV Project (Bo Peng et al., with EleutherAI and Recursal AI)","released":"2025-03-18","parameters":"0.19B to 2.9B (World v3 series); later 7.2B G1 checkpoint","parameters_billions":0.19,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/RWKV/RWKV7-Goose-World3-2.9B-HF","article_fact_checked":false},{"name":"ERNIE 4.5","article":"https://aiwiki.ai/wiki/ernie_4_5","kind":"language","developer":"Baidu","released":"2025-03-16","parameters":"0.3B dense up to 424B total (47B active) across ten variants","parameters_billions":0.3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/baidu/ERNIE-4.5-300B-A47B-PT","article_fact_checked":false},{"name":"Cohere Command A","article":"https://aiwiki.ai/wiki/command_a","kind":"language","developer":"Cohere","released":"2025-03-13","parameters":"111B","parameters_billions":111,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Gated on Hugging Face (contact details required). Commercial use only through Cohere's API, cloud partners or a Cohere license.","hugging_face":"https://huggingface.co/CohereLabs/c4ai-command-a-03-2025","article_fact_checked":false},{"name":"Sesame CSM","article":"https://aiwiki.ai/wiki/sesame_csm","kind":"audio","developer":"Sesame AI","released":"2025-03-13","parameters":"1B backbone + 100M audio decoder (CSM-1B)","parameters_billions":1,"tier":"partial","weights":"partial","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":false,"training_data_published":null,"note":"Only CSM-1B is released (Apache 2.0, click-through gate on Hugging Face); the larger 3B and 8B variants are not public.","hugging_face":"https://huggingface.co/sesame/csm-1b","article_fact_checked":false},{"name":"Gemma 3","article":"https://aiwiki.ai/wiki/gemma_3","kind":"language","developer":"Google DeepMind","released":"2025-03-12","parameters":"1B, 4B, 12B, 27B (270M added August 2025)","parameters_billions":1,"tier":"conditional","weights":"open","license":"gemma-terms","license_name":"Gemma Terms of Use","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Commercial use allowed under the Gemma Terms of Use and its prohibited use policy; Hugging Face access requires accepting the terms.","hugging_face":"https://huggingface.co/google/gemma-3-27b-it","article_fact_checked":true},{"name":"Skywork-R1V","article":"https://aiwiki.ai/wiki/skywork_r1v","kind":"language","developer":"Skywork AI (Kunlun Tech)","released":"2025-03","parameters":"38B (each open generation)","parameters_billions":38,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Skywork/Skywork-R1V3-38B","article_fact_checked":false},{"name":"Phi-4-mini","article":"https://aiwiki.ai/wiki/phi_4_mini","kind":"language","developer":"Microsoft","released":"2025-02-26","parameters":"3.8 billion","parameters_billions":3.8,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/microsoft/Phi-4-mini-instruct","article_fact_checked":false},{"name":"Wan 2.1","article":"https://aiwiki.ai/wiki/wan_2_1","kind":"video","developer":"Alibaba (Tongyi Wanxiang team)","released":"2025-02-25","parameters":"1.3B and 14B variants","parameters_billions":1.3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Wan-AI/Wan2.1-T2V-14B","article_fact_checked":true},{"name":"Evo 2 (genomic model)","article":"https://aiwiki.ai/wiki/evo_2","kind":"science","developer":"Arc Institute, NVIDIA, Stanford, and collaborators","released":"2025-02-19","parameters":"About 40 billion (largest); also 7B and 1B","parameters_billions":40,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/arcinstitute/evo2_40b","article_fact_checked":false},{"name":"Step-Video","article":"https://aiwiki.ai/wiki/step_video","kind":"video","developer":"StepFun (Shanghai Jieyue Xingchen)","released":"2025-02-17","parameters":"30B (DiT backbone)","parameters_billions":30,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/stepfun-ai/stepvideo-t2v","article_fact_checked":false},{"name":"LLaDA (Large Language Diffusion)","article":"https://aiwiki.ai/wiki/llada","kind":"language","developer":"Renmin University of China and Ant Group","released":"2025-02-14","parameters":"8B (LLaDA 8B); up to 100B (LLaDA 2.0 flash)","parameters_billions":8,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"LLaDA 8B Base/Instruct and LLaDA 1.5 are MIT; LLaDA 2.0 mini and flash (inclusionAI) are Apache 2.0.","hugging_face":"https://huggingface.co/GSAI-ML/LLaDA-8B-Base","article_fact_checked":false},{"name":"Goedel-Prover","article":"https://aiwiki.ai/wiki/goedel_prover","kind":"language","developer":"Princeton Language and Intelligence, with Tsinghua, Amazon, Meta FAIR, Numina","released":"2025-02","parameters":"7B (V1); 8B and 32B (V2)","parameters_billions":7,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"V2 weights (8B, 32B) are Apache 2.0; V1 weights are MIT.","hugging_face":"https://huggingface.co/Goedel-LM/Goedel-Prover-V2-32B","article_fact_checked":false},{"name":"NVIDIA COMPASS","article":"https://aiwiki.ai/wiki/nvidia_compass","kind":"robotics","developer":"NVIDIA","released":"2025-02","parameters":null,"parameters_billions":null,"tier":"conditional","weights":"open","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":true,"training_data_published":null,"note":"The checkpoint download requires accepting NVIDIA's terms on Hugging Face; the code is Apache 2.0.","hugging_face":"https://huggingface.co/nvidia/COMPASS","article_fact_checked":true},{"name":"olmOCR","article":"https://aiwiki.ai/wiki/olmocr","kind":"language","developer":"Allen Institute for AI (Ai2)","released":"2025-02","parameters":"7B","parameters_billions":7,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":true,"note":null,"hugging_face":"https://huggingface.co/allenai/olmOCR-2-7B-1025","article_fact_checked":false},{"name":"Qwen2.5-VL","article":"https://aiwiki.ai/wiki/qwen2_5_vl","kind":"vision","developer":"Alibaba Cloud (Qwen team)","released":"2025-01-26","parameters":"3B, 7B, 32B, 72B","parameters_billions":3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"7B and 32B are Apache 2.0. Exceptions: 3B is under the Qwen Research License (non-commercial); 72B is under the Qwen License.","hugging_face":null,"article_fact_checked":true},{"name":"DeepSeek-R1","article":"https://aiwiki.ai/wiki/deepseek_r1","kind":"language","developer":"DeepSeek","released":"2025-01-20","parameters":"671B total; 37B active per token","parameters_billions":671,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":false,"training_data_published":false,"note":"R1, R1-Zero and R1-0528 weights are MIT. Training data and the training framework were not released.","hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-R1","article_fact_checked":true},{"name":"DeepSeek-R1-Distill","article":"https://aiwiki.ai/wiki/deepseek_r1_distill","kind":"language","developer":"DeepSeek","released":"2025-01-20","parameters":"1.5B, 7B, 8B, 14B, 32B, and 70B variants","parameters_billions":1.5,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"All six are tagged MIT; the Qwen bases are Apache 2.0, and the two Llama-based variants also carry Llama 3.1 or 3.3 license terms.","hugging_face":null,"article_fact_checked":false},{"name":"MatterGen","article":"https://aiwiki.ai/wiki/mattergen","kind":"science","developer":"Microsoft Research, AI for Science","released":"2025-01-16","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/microsoft/mattergen","article_fact_checked":false},{"name":"MiniMax-Text-01","article":"https://aiwiki.ai/wiki/minimax_text_01","kind":"language","developer":"MiniMax","released":"2025-01-14","parameters":"456B total; 45.9B active per token","parameters_billions":456,"tier":"conditional","weights":"open","license":"custom","license_name":"MiniMax Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"License revised Feb 20, 2025. Commercial use is allowed subject to a prohibited-uses policy that bars military use.","hugging_face":"https://huggingface.co/MiniMaxAI/MiniMax-Text-01","article_fact_checked":false},{"name":"NVIDIA Cosmos","article":"https://aiwiki.ai/wiki/nvidia_cosmos","kind":"robotics","developer":"NVIDIA","released":"2025-01-06","parameters":"4B to 14B (Cosmos 1.0 models)","parameters_billions":4,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Cosmos 3 (2026) is under OpenMDW 1.1. Cosmos 1.0 to 2.5 stay under the NVIDIA Open Model License, with gated downloads.","hugging_face":null,"article_fact_checked":true},{"name":"DeepSeek V3","article":"https://aiwiki.ai/wiki/deepseek_v3","kind":"language","developer":"DeepSeek","released":"2024-12-26","parameters":"671B total (about 685B with MTP module); 37B active per token","parameters_billions":671,"tier":"conditional","weights":"open","license":"deepseek-model","license_name":"DeepSeek Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Original V3 and V3-Base weights use the DeepSeek Model License; the V3-0324 weights (March 2025) are MIT.","hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-V3","article_fact_checked":true},{"name":"QvQ","article":"https://aiwiki.ai/wiki/qvq","kind":"language","developer":"Qwen team, Alibaba","released":"2024-12-25","parameters":"72B (QVQ-72B-Preview)","parameters_billions":72,"tier":"partial","weights":"partial","license":"qwen-license","license_name":"Qwen License (Tongyi Qianwen)","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"QVQ-72B-Preview weights are under the Qwen license; QVQ-Max is offered only through Qwen Chat and Alibaba Cloud.","hugging_face":"https://huggingface.co/Qwen/QVQ-72B-Preview","article_fact_checked":false},{"name":"Falcon 3","article":"https://aiwiki.ai/wiki/falcon_3","kind":"language","developer":"Technology Innovation Institute (TII)","released":"2024-12-17","parameters":"1B, 3B, 7B, 10B (transformer); 7B (Mamba)","parameters_billions":1,"tier":"conditional","weights":"open","license":"falcon-llm","license_name":"TII Falcon License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Apache-2.0-derived license with an acceptable use policy; offering shared hosted instances needs a separate TII agreement.","hugging_face":"https://huggingface.co/tiiuae/Falcon3-7B-Base","article_fact_checked":false},{"name":"Byte Latent Transformer","article":"https://aiwiki.ai/wiki/byte_latent_transformer","kind":"language","developer":"Meta AI (FAIR)","released":"2024-12-13","parameters":"Up to 8B trained; 1B and 7B checkpoints released","parameters_billions":8,"tier":"noncommercial","weights":"open","license":"fair-noncommercial","license_name":"FAIR Noncommercial Research License","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"BLT 1B and 7B checkpoints are on Hugging Face behind an access request; the code is CC BY-NC 4.0.","hugging_face":"https://huggingface.co/facebook/blt-7b","article_fact_checked":false},{"name":"DeepSeek-VL2","article":"https://aiwiki.ai/wiki/deepseek_vl2","kind":"vision","developer":"DeepSeek (DeepSeek-AI)","released":"2024-12-13","parameters":"3B/16B/27B total; 1.0B/2.8B/4.5B active","parameters_billions":3,"tier":"conditional","weights":"open","license":"deepseek-model","license_name":"DeepSeek Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Code is MIT; weights use the DeepSeek Model License, which allows commercial use subject to use restrictions.","hugging_face":"https://huggingface.co/deepseek-ai/deepseek-vl2","article_fact_checked":false},{"name":"Meta Motivo","article":"https://aiwiki.ai/wiki/meta_motivo","kind":"robotics","developer":"Meta AI (FAIR)","released":"2024-12-12","parameters":"~24.5M (S variants); ~288M (metamotivo-M-1)","parameters_billions":0.0245,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/facebook/metamotivo-M-1","article_fact_checked":false},{"name":"Phi-4","article":"https://aiwiki.ai/wiki/phi_4","kind":"language","developer":"Microsoft Research","released":"2024-12-12","parameters":"14 billion (about 14.7B with embeddings)","parameters_billions":14,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/microsoft/phi-4","article_fact_checked":true},{"name":"Llama 3.3","article":"https://aiwiki.ai/wiki/llama_3_3","kind":"language","developer":"Meta","released":"2024-12-06","parameters":"70 billion","parameters_billions":70,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated download. Acceptable use policy applies; over 700M monthly active users needs a separate Meta license.","hugging_face":"https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct","article_fact_checked":true},{"name":"GenCast","article":"https://aiwiki.ai/wiki/gencast","kind":"science","developer":"Google DeepMind","released":"2024-12-04","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"cc-by-4.0","license_name":"CC BY 4.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Weights were relicensed from CC BY-NC-SA 4.0 to CC BY 4.0 on August 6, 2026 to permit commercial use; code is Apache 2.0.","hugging_face":null,"article_fact_checked":false},{"name":"HunyuanVideo","article":"https://aiwiki.ai/wiki/hunyuan_video","kind":"video","developer":"Tencent","released":"2024-12-03","parameters":"13B (1.0); 8.3B (1.5)","parameters_billions":13,"tier":"conditional","weights":"open","license":"hunyuan-community","license_name":"Tencent Hunyuan Community License","commercial_use":"conditional","training_code_published":true,"training_data_published":null,"note":"License does not apply in the EU, UK or South Korea; over 100M monthly active users needs a separate Tencent license.","hugging_face":"https://huggingface.co/tencent/HunyuanVideo","article_fact_checked":true},{"name":"QwQ","article":"https://aiwiki.ai/wiki/qwq","kind":"language","developer":"Alibaba Cloud Qwen team","released":"2024-11-28","parameters":"32.5B (dense)","parameters_billions":32.5,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/QwQ-32B","article_fact_checked":false},{"name":"OLMo 2","article":"https://aiwiki.ai/wiki/olmo_2","kind":"language","developer":"Allen Institute for AI (Ai2)","released":"2024-11-26","parameters":"7B, 13B, 32B","parameters_billions":7,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/allenai/OLMo-2-0325-32B","article_fact_checked":true},{"name":"LTX-Video","article":"https://aiwiki.ai/wiki/ltx_video","kind":"video","developer":"Lightricks","released":"2024-11-21","parameters":"~1.9B (base); 13B (LTXV-13B)","parameters_billions":1.9,"tier":"conditional","weights":"open","license":"custom","license_name":"LTXV Open Weights License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Entities with annual revenue of $10M or more need a paid commercial license. Versions before 0.9.6 shipped under earlier licenses (0.9: RAIL-M).","hugging_face":"https://huggingface.co/Lightricks/LTX-Video","article_fact_checked":false},{"name":"Tülu 3","article":"https://aiwiki.ai/wiki/tulu_3","kind":"other","developer":"Allen Institute for AI (Ai2)","released":"2024-11-21","parameters":"8B, 70B, 405B","parameters_billions":8,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":true,"training_data_published":true,"note":"Inherits the Llama 3.1 Community License (acceptable use policy, 700M monthly user threshold); the card also cites Gemma and Qwen terms.","hugging_face":null,"article_fact_checked":true},{"name":"Pixtral Large","article":"https://aiwiki.ai/wiki/pixtral_large","kind":"language","developer":"Mistral AI","released":"2024-11-18","parameters":"~124 billion total","parameters_billions":124,"tier":"noncommercial","weights":"open","license":"mistral-research","license_name":"Mistral Research License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Research and non-commercial use only; commercial use needs a separate Mistral Commercial License.","hugging_face":"https://huggingface.co/mistralai/Pixtral-Large-Instruct-2411","article_fact_checked":false},{"name":"Boltz","article":"https://aiwiki.ai/wiki/boltz","kind":"science","developer":"MIT (CSAIL and Jameel Clinic)","released":"2024-11-01","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":false},{"name":"SmolLM 2","article":"https://aiwiki.ai/wiki/smollm_2","kind":"language","developer":"Hugging Face","released":"2024-11-01","parameters":"135M, 360M, 1.7B (three sizes)","parameters_billions":0.135,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B","article_fact_checked":false},{"name":"Hunyuan 3D","article":"https://aiwiki.ai/wiki/hunyuan_3d","kind":"other","developer":"Tencent","released":"2024-11","parameters":"Up to ~10B shape model (2.5); 3.3B shape + 2B paint in 2.1","parameters_billions":10,"tier":"partial","weights":"partial","license":"hunyuan-community","license_name":"Tencent Hunyuan Community License","commercial_use":"conditional","training_code_published":true,"training_data_published":null,"note":"Weights open up to 2.1; 2.5, 3.0 and 3.1 are hosted only. Excludes EU, UK and South Korea; above 1M monthly users needs a separate license.","hugging_face":"https://huggingface.co/tencent/Hunyuan3D-2","article_fact_checked":true},{"name":"Marco-o1","article":"https://aiwiki.ai/wiki/marco_o1","kind":"language","developer":"Alibaba International Digital Commerce (MarcoPolo team)","released":"2024-11","parameters":"About 7.62B (7,615,616,512 in the published checkpoint)","parameters_billions":7.62,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/ATH-MaaS/Marco-o1","article_fact_checked":false},{"name":"π0","article":"https://aiwiki.ai/wiki/pi0","kind":"robotics","developer":"Physical Intelligence","released":"2024-10-31","parameters":"~3.3B total (3B VLM backbone + ~300M action expert)","parameters_billions":3.3,"tier":"conditional","weights":"open","license":"gemma-terms","license_name":"Gemma Terms of Use","commercial_use":"conditional","training_code_published":true,"training_data_published":false,"note":"openpi is Apache 2.0 and also ships the Gemma Terms of Use, which cover the PaliGemma-derived weights. The pi training dataset is proprietary.","hugging_face":null,"article_fact_checked":true},{"name":"GLM-4-Voice","article":"https://aiwiki.ai/wiki/glm_4_voice","kind":"audio","developer":"Zhipu AI and Tsinghua University KEG","released":"2024-10-25","parameters":"9B (GLM-4-9B backbone)","parameters_billions":9,"tier":"conditional","weights":"open","license":"custom","license_name":"glm-4-voice License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Free for academic research; commercial use requires registration with Zhipu AI and a Built with glm-4 notice. Code is Apache 2.0.","hugging_face":"https://huggingface.co/zai-org/glm-4-voice-9b","article_fact_checked":false},{"name":"Mochi 1","article":"https://aiwiki.ai/wiki/mochi_1","kind":"video","developer":"Genmo Inc.","released":"2024-10-22","parameters":"10 billion (DiT) plus 362 million (VAE)","parameters_billions":10,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/genmo/mochi-1-preview","article_fact_checked":false},{"name":"Stable Diffusion 3.5","article":"https://aiwiki.ai/wiki/stable_diffusion_3_5","kind":"image","developer":"Stability AI","released":"2024-10-22","parameters":"8.1B (Large); ~2.5B (Medium)","parameters_billions":8.1,"tier":"conditional","weights":"open","license":"stability-community","license_name":"Stability AI Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Free for commercial use under $1M annual revenue, enterprise license above that; click-through gate on Hugging Face.","hugging_face":"https://huggingface.co/stabilityai/stable-diffusion-3.5-large","article_fact_checked":false},{"name":"SpiRit-LM","article":"https://aiwiki.ai/wiki/spirit_lm","kind":"audio","developer":"Meta AI (FAIR)","released":"2024-10-18","parameters":"7B (Base and Expressive versions)","parameters_billions":7,"tier":"noncommercial","weights":"open","license":"fair-noncommercial","license_name":"FAIR Noncommercial Research License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Base and Expressive 7B weights are released for noncommercial research only, with inference code.","hugging_face":null,"article_fact_checked":false},{"name":"DeepSeek Janus","article":"https://aiwiki.ai/wiki/deepseek_janus","kind":"vision","developer":"DeepSeek","released":"2024-10-17","parameters":"1.3B (Janus, JanusFlow); 1B and 7B (Janus-Pro)","parameters_billions":1.3,"tier":"conditional","weights":"open","license":"deepseek-model","license_name":"DeepSeek Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Hub tag reads MIT, but the model cards put the weights under the DeepSeek Model License; only the code is MIT.","hugging_face":"https://huggingface.co/deepseek-ai/Janus-Pro-7B","article_fact_checked":false},{"name":"Ministral 3B / 8B","article":"https://aiwiki.ai/wiki/ministral","kind":"language","developer":"Mistral AI","released":"2024-10-16","parameters":"~3B and ~8B (8,019,808,256 for the 8B)","parameters_billions":3,"tier":"partial","weights":"partial","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Ministral 3 (3B, 8B, 14B) is Apache 2.0. Ministral 8B 2410 is under the Mistral Research License and Ministral 3B 2410 was API only.","hugging_face":"https://huggingface.co/mistralai/Ministral-3-8B-Instruct-2512","article_fact_checked":false},{"name":"F5-TTS","article":"https://aiwiki.ai/wiki/f5_tts","kind":"audio","developer":"Shanghai Jiao Tong University X-LANCE Lab, University of Cambridge, Geely","released":"2024-10-09","parameters":"335.8 million","parameters_billions":0.33580000000000004,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":true,"note":"Code is MIT; the official checkpoints are CC BY-NC 4.0 because of the Emilia training data.","hugging_face":"https://huggingface.co/SWivid/F5-TTS","article_fact_checked":false},{"name":"Llama-3.1-Nemotron-70B-Instruct","article":"https://aiwiki.ai/wiki/llama_3_1_nemotron_70b","kind":"language","developer":"NVIDIA","released":"2024-10","parameters":"70B","parameters_billions":70,"tier":"conditional","weights":"open","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"The Llama 3.1 Community License also applies, and the Hugging Face tag is llama3.1.","hugging_face":"https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Instruct-HF","article_fact_checked":false},{"name":"Llama 3.2","article":"https://aiwiki.ai/wiki/llama_3_2","kind":"language","developer":"Meta","released":"2024-09-25","parameters":"1.23B, 3.21B (text-only); 11B, 90B (vision)","parameters_billions":1.23,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated download. For the 11B and 90B vision models the license grant excludes individuals and companies based in the EU.","hugging_face":null,"article_fact_checked":true},{"name":"Llama 3.2 Vision","article":"https://aiwiki.ai/wiki/llama_3_2_vision","kind":"language","developer":"Meta AI","released":"2024-09-25","parameters":"11B (10.6B actual) and 90B (88.8B actual)","parameters_billions":11,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated download. The license grant excludes individuals and companies based in the EU (end users of products are exempt).","hugging_face":"https://huggingface.co/meta-llama/Llama-3.2-90B-Vision-Instruct","article_fact_checked":true},{"name":"Molmo","article":"https://aiwiki.ai/wiki/molmo","kind":"vision","developer":"Allen Institute for AI (Ai2)","released":"2024-09-25","parameters":"72B, 7B (x2), MolmoE-1B (~1.5B active, 7.2B total)","parameters_billions":72,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":true,"note":"Training code is on GitHub and the PixMo training datasets are published under ODC-BY 1.0.","hugging_face":"https://huggingface.co/allenai/Molmo-72B-0924","article_fact_checked":false},{"name":"Qwen2.5","article":"https://aiwiki.ai/wiki/qwen2_5","kind":"language","developer":"Alibaba Cloud Qwen team","released":"2024-09-19","parameters":"Seven dense sizes from 0.49B to 72.7B","parameters_billions":0.49,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Five of seven sizes are Apache 2.0; the 72B is under the Qwen License (conditions on commercial use) and the 3B is research only.","hugging_face":"https://huggingface.co/Qwen/Qwen2.5-7B-Instruct","article_fact_checked":true},{"name":"Qwen2.5-Math","article":"https://aiwiki.ai/wiki/qwen2_5_math","kind":"language","developer":"Qwen team, Alibaba Cloud","released":"2024-09-19","parameters":"1.5B, 7B and 72B","parameters_billions":1.5,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"The 1.5B and 7B models are Apache 2.0; the 72B models and RM-72B are under the Qwen License (conditions on commercial use).","hugging_face":"https://huggingface.co/Qwen/Qwen2.5-Math-7B-Instruct","article_fact_checked":true},{"name":"Jina Embeddings v3","article":"https://aiwiki.ai/wiki/jina_embeddings_v3","kind":"language","developer":"Jina AI","released":"2024-09-18","parameters":"570 million","parameters_billions":0.5700000000000001,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/jinaai/jina-embeddings-v3","article_fact_checked":false},{"name":"NVLM","article":"https://aiwiki.ai/wiki/nvlm","kind":"language","developer":"NVIDIA","released":"2024-09-17","parameters":"~79B (NVLM-D-72B: 72B LLM plus 6B vision encoder)","parameters_billions":79,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/nvidia/NVLM-D-72B","article_fact_checked":false},{"name":"Pixtral","article":"https://aiwiki.ai/wiki/pixtral","kind":"language","developer":"Mistral AI","released":"2024-09-11","parameters":"~12.4B (12B: 12B decoder + 400M encoder); 124B (Large)","parameters_billions":12.4,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Applies to Pixtral 12B. Pixtral Large (124B) is under the Mistral Research License: non-commercial unless a commercial license is bought.","hugging_face":null,"article_fact_checked":true},{"name":"Chai-1","article":"https://aiwiki.ai/wiki/chai_1","kind":"science","developer":"Chai Discovery","released":"2024-09-09","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":false,"training_data_published":null,"note":"Apache 2.0 for code and weights since November 26, 2024 (research-only license before that). Inference code only; no training code.","hugging_face":"https://huggingface.co/chaidiscovery/chai-1","article_fact_checked":false},{"name":"OLMoE","article":"https://aiwiki.ai/wiki/olmoe","kind":"language","developer":"Allen Institute for AI, with Contextual AI","released":"2024-09-03","parameters":"6.9B total; ~1.3B active per token","parameters_billions":6.9,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/allenai/OLMoE-1B-7B-0924","article_fact_checked":false},{"name":"Hailuo AI","article":"https://aiwiki.ai/wiki/hailuo","kind":"video","developer":"MiniMax","released":"2024-09","parameters":null,"parameters_billions":null,"tier":"partial","weights":"partial","license":"custom","license_name":"MiniMax H3 Community License Agreement","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Only MiniMax H3 weights are public (Aug 2026), in part; earlier Hailuo models are hosted only. License excludes the EU, UK, South Korea and US.","hugging_face":null,"article_fact_checked":true},{"name":"Qwen2.5-Coder","article":"https://aiwiki.ai/wiki/qwen2_5_coder","kind":"language","developer":"Qwen team, Alibaba","released":"2024-09","parameters":"Six sizes, 0.5B to 32B (flagship 32.5B)","parameters_billions":0.5,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Five sizes (0.5B, 1.5B, 7B, 14B, 32B) are Apache 2.0; the 3B is under the Qwen Research License, which is non-commercial.","hugging_face":"https://huggingface.co/Qwen/Qwen2.5-Coder-32B-Instruct","article_fact_checked":true},{"name":"Qwen2-VL","article":"https://aiwiki.ai/wiki/qwen2_vl","kind":"vision","developer":"Alibaba Cloud (Qwen team)","released":"2024-08-29","parameters":"2B, 7B, 72B","parameters_billions":2,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"2B and 7B are Apache 2.0. Exception: the 72B weights use the Qwen (Tongyi Qianwen) License, which sets conditions on commercial use.","hugging_face":null,"article_fact_checked":true},{"name":"Sapiens (computer vision)","article":"https://aiwiki.ai/wiki/sapiens","kind":"vision","developer":"Meta (Reality Labs)","released":"2024-08-22","parameters":"0.3B, 0.6B, 1B, 2B","parameters_billions":0.3,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Code and weights of the 2024 release are CC BY-NC 4.0; the 2026 successor Sapiens2 uses its own Sapiens2 License.","hugging_face":null,"article_fact_checked":false},{"name":"Show-o","article":"https://aiwiki.ai/wiki/show_o","kind":"other","developer":"Show Lab, National University of Singapore (with ByteDance co-authors)","released":"2024-08-21","parameters":"About 1.3B","parameters_billions":1.3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"GitHub LICENSE and the 512x512 and Show-o2 checkpoints are Apache 2.0; the first 256x256 Hub checkpoints are tagged MIT.","hugging_face":null,"article_fact_checked":false},{"name":"Qwen2-Math","article":"https://aiwiki.ai/wiki/qwen2_math","kind":"language","developer":"Qwen team, Alibaba","released":"2024-08-08","parameters":"1.5B, 7B, and 72B (base and Instruct)","parameters_billions":1.5,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"The 1.5B and 7B sizes are Apache 2.0; Qwen2-Math-72B is under the Tongyi Qianwen License (conditions on commercial use).","hugging_face":"https://huggingface.co/Qwen/Qwen2-Math-7B-Instruct","article_fact_checked":false},{"name":"CogVideoX","article":"https://aiwiki.ai/wiki/cogvideo","kind":"video","developer":"Zhipu AI and Tsinghua University (THUDM)","released":"2024-08-06","parameters":"2B or 5B (by variant)","parameters_billions":2,"tier":"conditional","weights":"open","license":"custom","license_name":"CogVideoX License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"CogVideoX-2B is Apache 2.0; 5B and 1.5 use the CogVideoX License: commercial use needs free registration, capped at 1M visits a month.","hugging_face":"https://huggingface.co/zai-org/CogVideoX-5b","article_fact_checked":false},{"name":"Flux (text-to-image model)","article":"https://aiwiki.ai/wiki/flux","kind":"image","developer":"Black Forest Labs","released":"2024-08-01","parameters":"12B (FLUX.1); 32B (FLUX.2); 4B and 9B (FLUX.2 klein)","parameters_billions":12,"tier":"partial","weights":"partial","license":"flux-non-commercial","license_name":"FLUX Non-Commercial License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"[schnell] and FLUX.2 klein 4B are Apache 2.0; [dev] and klein 9B are non-commercial; FLUX 3 Klein and Action use the FLUX Kommunity License; pro tiers API only.","hugging_face":null,"article_fact_checked":true},{"name":"FLUX.1","article":"https://aiwiki.ai/wiki/flux_1","kind":"image","developer":"Black Forest Labs","released":"2024-08-01","parameters":"~12B","parameters_billions":12,"tier":"partial","weights":"partial","license":"flux-non-commercial","license_name":"FLUX Non-Commercial License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"[schnell] is Apache 2.0; [dev], Kontext [dev], Krea [dev] and the Tools are non-commercial; [pro] is API only. BFL sells commercial licenses.","hugging_face":"https://huggingface.co/black-forest-labs/FLUX.1-dev","article_fact_checked":true},{"name":"Qwen2-Audio","article":"https://aiwiki.ai/wiki/qwen2_audio","kind":"audio","developer":"Qwen team, Alibaba Cloud","released":"2024-08","parameters":"About 8.2B","parameters_billions":8.2,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Qwen/Qwen2-Audio-7B-Instruct","article_fact_checked":false},{"name":"ShieldGemma","article":"https://aiwiki.ai/wiki/shieldgemma","kind":"other","developer":"Google","released":"2024-07-31","parameters":"2B, 9B, 27B (text); 4B (ShieldGemma 2)","parameters_billions":2,"tier":"conditional","weights":"open","license":"gemma-terms","license_name":"Gemma Terms of Use","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated: Google's Gemma terms must be accepted before download. Covers ShieldGemma and ShieldGemma 2.","hugging_face":null,"article_fact_checked":false},{"name":"SAM 2","article":"https://aiwiki.ai/wiki/sam_2","kind":"vision","developer":"Meta AI (FAIR)","released":"2024-07-29","parameters":"Tiny 38.9M, Small 46M, Base+ 80.8M, Large 224.4M","parameters_billions":0.0389,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/facebook/sam2.1-hiera-large","article_fact_checked":false},{"name":"Llama 3.1","article":"https://aiwiki.ai/wiki/llama_3_1","kind":"language","developer":"Meta","released":"2024-07-23","parameters":"8B, 70B, and 405B","parameters_billions":8,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated download. Acceptable use policy applies; over 700M monthly active users needs a separate Meta license.","hugging_face":"https://huggingface.co/meta-llama/Llama-3.1-405B","article_fact_checked":true},{"name":"Mistral NeMo","article":"https://aiwiki.ai/wiki/mistral_nemo","kind":"language","developer":"Mistral AI, with NVIDIA","released":"2024-07-18","parameters":"12B","parameters_billions":12,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/mistralai/Mistral-Nemo-Instruct-2407","article_fact_checked":true},{"name":"SmolLM","article":"https://aiwiki.ai/wiki/smollm","kind":"language","developer":"Hugging Face","released":"2024-07-16","parameters":"135M, 360M, and 1.7B variants","parameters_billions":0.135,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/HuggingFaceTB/SmolLM-1.7B","article_fact_checked":false},{"name":"Moshi","article":"https://aiwiki.ai/wiki/moshi","kind":"audio","developer":"Kyutai","released":"2024-07-03","parameters":"7.7B total (Helium backbone is 7B)","parameters_billions":7.7,"tier":"permissive","weights":"open","license":"cc-by-4.0","license_name":"CC BY 4.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/kyutai/moshiko-pytorch-bf16","article_fact_checked":false},{"name":"CosyVoice","article":"https://aiwiki.ai/wiki/cosyvoice","kind":"audio","developer":"Tongyi SpeechTeam, Alibaba Group (FunAudioLLM project)","released":"2024-07","parameters":"300M (v1); 0.5B (CosyVoice 2 and the open CosyVoice 3 checkpoint); 1.5B language model in the CosyVoice 3 paper","parameters_billions":0.3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":false},{"name":"Gemma 2","article":"https://aiwiki.ai/wiki/gemma_2","kind":"language","developer":"Google DeepMind","released":"2024-06-27","parameters":"2.6B, 9B, 27B","parameters_billions":2.6,"tier":"conditional","weights":"open","license":"gemma-terms","license_name":"Gemma Terms of Use","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Commercial use allowed under the Gemma Terms of Use and its prohibited use policy; Hugging Face access requires accepting the terms.","hugging_face":"https://huggingface.co/google/gemma-2-27b","article_fact_checked":true},{"name":"LLM Compiler (Meta)","article":"https://aiwiki.ai/wiki/llm_compiler","kind":"language","developer":"Meta AI","released":"2024-06-27","parameters":"7B and 13B (base and FTD variants)","parameters_billions":7,"tier":"conditional","weights":"open","license":"custom","license_name":"Meta LLM Compiler License Agreement","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated download. Llama acceptable use policy applies; over 700M monthly active users needs a separate Meta license.","hugging_face":"https://huggingface.co/facebook/llm-compiler-13b","article_fact_checked":false},{"name":"ESM3","article":"https://aiwiki.ai/wiki/esm3","kind":"science","developer":"EvolutionaryScale","released":"2024-06-25","parameters":"98B flagship; also 1.4B and 7B","parameters_billions":98,"tier":"partial","weights":"partial","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Only the 1.4B esm3-sm-open-v1 is downloadable; 7B and 98B are API only. Relicensed to MIT by Biohub in 2026 (was Cambrian Non-Commercial).","hugging_face":"https://huggingface.co/biohub/esm3-sm-open-v1","article_fact_checked":false},{"name":"Chameleon (Meta AI)","article":"https://aiwiki.ai/wiki/chameleon","kind":"vision","developer":"Meta AI (FAIR)","released":"2024-06-18","parameters":"7B and 34B","parameters_billions":7,"tier":"noncommercial","weights":"open","license":"custom","license_name":"Chameleon Research License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"7B and 34B weights are released for research use only, after an access request.","hugging_face":"https://huggingface.co/facebook/chameleon-30b","article_fact_checked":false},{"name":"Florence-2","article":"https://aiwiki.ai/wiki/florence_2","kind":"vision","developer":"Microsoft Research","released":"2024-06-16","parameters":"232M (base); 771M (large)","parameters_billions":0.232,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/microsoft/Florence-2-large","article_fact_checked":false},{"name":"Stable Diffusion 3","article":"https://aiwiki.ai/wiki/stable_diffusion_3","kind":"image","developer":"Stability AI","released":"2024-06-12","parameters":"800M to 8.1B (SD 3.5 Large)","parameters_billions":0.8,"tier":"partial","weights":"partial","license":"stability-community","license_name":"Stability AI Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"SD3 Medium and all SD 3.5 models are downloadable; SD3 Large and Large Turbo were API only. Free commercial use under $1M annual revenue.","hugging_face":"https://huggingface.co/stabilityai/stable-diffusion-3-medium","article_fact_checked":true},{"name":"Qwen2","article":"https://aiwiki.ai/wiki/qwen2","kind":"language","developer":"Alibaba Cloud","released":"2024-06-06","parameters":"0.5B, 1.5B, 7B, 57B-A14B (14B active), 72B","parameters_billions":0.5,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"The 0.5B, 1.5B, 7B and 57B-A14B sizes are Apache 2.0; Qwen2-72B is under the Tongyi Qianwen License (conditions on commercial use).","hugging_face":"https://huggingface.co/Qwen/Qwen2-7B-Instruct","article_fact_checked":true},{"name":"OpenVLA","article":"https://aiwiki.ai/wiki/openvla","kind":"robotics","developer":"Stanford, UC Berkeley, TRI, Google DeepMind, Physical Intelligence, MIT","released":"2024-06","parameters":"7B","parameters_billions":7,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":true,"training_data_published":true,"note":"Code is MIT and the Hub tag reads MIT, but the developers say the checkpoints derive from Llama 2 and fall under the Llama Community License.","hugging_face":"https://huggingface.co/openvla/openvla-7b","article_fact_checked":true},{"name":"Codestral","article":"https://aiwiki.ai/wiki/codestral","kind":"language","developer":"Mistral AI","released":"2024-05-29","parameters":"22.2B (22B v0.1); 7.3B (Mamba); later versions not disclosed","parameters_billions":22.2,"tier":"partial","weights":"partial","license":"mnpl","license_name":"Mistral Non-Production License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Codestral 22B weights are MNPL (non-production); Codestral Mamba 7B is Apache 2.0; Codestral 25.01, 25.08 and Embed are API only.","hugging_face":"https://huggingface.co/mistralai/Codestral-22B-v0.1","article_fact_checked":true},{"name":"DeepSeek-Prover","article":"https://aiwiki.ai/wiki/deepseek_prover","kind":"language","developer":"DeepSeek","released":"2024-05-23","parameters":"7B; 671B MoE (V2)","parameters_billions":7,"tier":"conditional","weights":"open","license":"deepseek-model","license_name":"DeepSeek Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"V1, V1.5 and V2 weights use the DeepSeek License; the V2 Hugging Face repos have no license tag and ship an MIT LICENSE file.","hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-Prover-V2-671B","article_fact_checked":false},{"name":"PaliGemma","article":"https://aiwiki.ai/wiki/paligemma","kind":"vision","developer":"Google","released":"2024-05-14","parameters":"3B (PaliGemma); 3B, 10B and 28B (PaliGemma 2)","parameters_billions":3,"tier":"conditional","weights":"open","license":"gemma-terms","license_name":"Gemma Terms of Use","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated on Hugging Face: the Gemma terms must be accepted before download. Covers PaliGemma and PaliGemma 2.","hugging_face":null,"article_fact_checked":true},{"name":"AlphaFold 3","article":"https://aiwiki.ai/wiki/alphafold_3","kind":"science","developer":"Google DeepMind and Isomorphic Labs","released":"2024-05-08","parameters":null,"parameters_billions":null,"tier":"noncommercial","weights":"open","license":"custom","license_name":"AlphaFold 3 Model Parameters Terms of Use","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Weights are for non-commercial use by non-commercial organizations only and may not be redistributed. The code is Apache 2.0 since June 2026.","hugging_face":null,"article_fact_checked":true},{"name":"DeepSeek-V2","article":"https://aiwiki.ai/wiki/deepseek_v2","kind":"language","developer":"DeepSeek","released":"2024-05","parameters":"236B total; ~21B active per token","parameters_billions":236,"tier":"conditional","weights":"open","license":"deepseek-model","license_name":"DeepSeek Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Code is MIT; weights use the DeepSeek License, which allows commercial use subject to use-based restrictions.","hugging_face":"https://huggingface.co/deepseek-ai/DeepSeek-V2","article_fact_checked":true},{"name":"IBM Granite","article":"https://aiwiki.ai/wiki/ibm_granite","kind":"language","developer":"IBM","released":"2024-05","parameters":"~350M to 32B MoE (H-Small, ~9B active)","parameters_billions":0.35000000000000003,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":false,"note":"The first 2023 Granite models were proprietary (watsonx only); releases since Granite Code (May 2024) are Apache 2.0.","hugging_face":null,"article_fact_checked":false},{"name":"Snowflake Arctic","article":"https://aiwiki.ai/wiki/snowflake_arctic","kind":"language","developer":"Snowflake AI Research","released":"2024-04-24","parameters":"480B total; 17B active per token","parameters_billions":480,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/Snowflake/snowflake-arctic-instruct","article_fact_checked":false},{"name":"Phi-3","article":"https://aiwiki.ai/wiki/phi_3","kind":"language","developer":"Microsoft","released":"2024-04-23","parameters":"3.8B (Mini) to 14B (Medium); Phi-3.5-MoE 42B total, 6.6B active","parameters_billions":3.8,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/microsoft/Phi-3-mini-4k-instruct","article_fact_checked":true},{"name":"Llama 3","article":"https://aiwiki.ai/wiki/llama_3","kind":"language","developer":"Meta","released":"2024-04-18","parameters":"8B and 70B (base and Instruct variants)","parameters_billions":8,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated download. Acceptable use policy applies; over 700M monthly active users needs a separate Meta license.","hugging_face":"https://huggingface.co/meta-llama/Meta-Llama-3-8B","article_fact_checked":true},{"name":"Mixtral 8x22B","article":"https://aiwiki.ai/wiki/mixtral_8x22b","kind":"language","developer":"Mistral AI","released":"2024-04-17","parameters":"~141B total; ~39B active per token","parameters_billions":141,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/mistralai/Mixtral-8x22B-Instruct-v0.1","article_fact_checked":true},{"name":"RecurrentGemma","article":"https://aiwiki.ai/wiki/recurrentgemma","kind":"language","developer":"Google DeepMind","released":"2024-04-11","parameters":"2.68B and 8.58B total","parameters_billions":2.68,"tier":"conditional","weights":"open","license":"gemma-terms","license_name":"Gemma Terms of Use","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated on Hugging Face and Kaggle: accept the Gemma Terms of Use first. The reference code is Apache 2.0.","hugging_face":"https://huggingface.co/google/recurrentgemma-9b","article_fact_checked":false},{"name":"CodeGemma","article":"https://aiwiki.ai/wiki/codegemma","kind":"language","developer":"Google","released":"2024-04-09","parameters":"2B to 7B","parameters_billions":2,"tier":"conditional","weights":"open","license":"gemma-terms","license_name":"Gemma Terms of Use","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Download requires accepting the Gemma Terms of Use, which include a prohibited-use policy.","hugging_face":null,"article_fact_checked":false},{"name":"Command R+","article":"https://aiwiki.ai/wiki/command_r_plus","kind":"language","developer":"Cohere","released":"2024-04-04","parameters":"104 billion","parameters_billions":104,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Gated on Hugging Face (contact details required); Cohere Labs Acceptable Use Policy also applies. Commercial use only through Cohere.","hugging_face":"https://huggingface.co/CohereLabs/c4ai-command-r-plus","article_fact_checked":false},{"name":"NVIDIA Canary","article":"https://aiwiki.ai/wiki/canary","kind":"audio","developer":"NVIDIA","released":"2024-04","parameters":"182M to 2.5B across variants","parameters_billions":0.182,"tier":"permissive","weights":"open","license":"cc-by-4.0","license_name":"CC BY 4.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"The 2025 models (180M-Flash, 1B-Flash, 1B-v2, Qwen-2.5B) are CC BY 4.0; the original Canary-1B (April 2024) is CC BY-NC 4.0, non-commercial.","hugging_face":null,"article_fact_checked":false},{"name":"Reka Edge","article":"https://aiwiki.ai/wiki/reka_edge","kind":"language","developer":"Reka AI","released":"2024-04","parameters":"7 billion","parameters_billions":7,"tier":"partial","weights":"partial","license":"custom","license_name":"Reka Edge 2603 License (Business Source License 1.1)","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Only the 2026 redesign (reka-edge-2603) is downloadable; production use is free under $1M annual revenue. The 2024 Reka Edge was never released as weights.","hugging_face":null,"article_fact_checked":false},{"name":"Jamba","article":"https://aiwiki.ai/wiki/jamba","kind":"language","developer":"AI21 Labs","released":"2024-03-28","parameters":"52B total; 12B active (Jamba 1.0); up to 398B total (Jamba 1.5 Large)","parameters_billions":52,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Jamba 1.0, Jamba Reasoning 3B and Jamba 2 are Apache 2.0; Jamba 1.5, 1.6 and 1.7 use the Jamba Open Model License.","hugging_face":"https://huggingface.co/ai21labs/Jamba-v0.1","article_fact_checked":false},{"name":"DBRX","article":"https://aiwiki.ai/wiki/dbrx","kind":"language","developer":"Databricks (Mosaic AI)","released":"2024-03-27","parameters":"132B total; 36B active per token","parameters_billions":132,"tier":"conditional","weights":"open","license":"custom","license_name":"Databricks Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Official Databricks Hugging Face and GitHub repos were offline in Nov 2026; mirrors keep the same license. Separate license above 700M MAU.","hugging_face":null,"article_fact_checked":false},{"name":"NVIDIA Earth-2","article":"https://aiwiki.ai/wiki/nvidia_earth_2","kind":"science","developer":"NVIDIA","released":"2024-03-18","parameters":null,"parameters_billions":null,"tier":"conditional","weights":"open","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Varies by model: Atlas, StormScope and the first HealDA use the NVIDIA Open Model License; FourCastNet 3 and cBottle are Apache 2.0; newer ones OpenMDW 1.1.","hugging_face":null,"article_fact_checked":false},{"name":"Open-Sora","article":"https://aiwiki.ai/wiki/open_sora","kind":"video","developer":"HPC-AI Tech (Colossal-AI team)","released":"2024-03-17","parameters":"11B (v2.0)","parameters_billions":11,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/hpcai-tech/Open-Sora-v2","article_fact_checked":false},{"name":"Command R","article":"https://aiwiki.ai/wiki/command_r","kind":"language","developer":"Cohere","released":"2024-03-11","parameters":"35B (Command R); family spans 7B to 104B","parameters_billions":35,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"All open members (Command R v01 and 08-2024, R7B, R+) are CC BY-NC 4.0 and gated on Hugging Face (contact details required).","hugging_face":"https://huggingface.co/CohereLabs/c4ai-command-r-v01","article_fact_checked":false},{"name":"DeepSeek-VL","article":"https://aiwiki.ai/wiki/deepseek_vl","kind":"vision","developer":"DeepSeek","released":"2024-03-11","parameters":"1.3B and 7B","parameters_billions":1.3,"tier":"conditional","weights":"open","license":"deepseek-model","license_name":"DeepSeek Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Code is MIT; weights use the DeepSeek Model License, which allows commercial use subject to use restrictions.","hugging_face":"https://huggingface.co/deepseek-ai/deepseek-vl-7b-chat","article_fact_checked":false},{"name":"BitNet b1.58","article":"https://aiwiki.ai/wiki/bitnet_b1_58","kind":"language","developer":"Microsoft Research","released":"2024-02-27","parameters":"700M to 3B (original); 2B (BitNet b1.58 2B4T)","parameters_billions":0.7000000000000001,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"The official open weights are BitNet b1.58 2B4T (April 2025), under MIT.","hugging_face":"https://huggingface.co/microsoft/bitnet-b1.58-2B-4T","article_fact_checked":false},{"name":"Mistral Large","article":"https://aiwiki.ai/wiki/mistral_large","kind":"language","developer":"Mistral AI","released":"2024-02-26","parameters":"123B (Large 2/2.1); 675B total, 41B active (Large 3)","parameters_billions":123,"tier":"partial","weights":"partial","license":"custom","license_name":"Mistral Ecosystem Open-Weight License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Large 4: Model-as-a-Service operators need a commercial license. Large 3 is Apache 2.0; Large 2 and 2.1 are research only; Large 1 was API only.","hugging_face":"https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B","article_fact_checked":true},{"name":"Nemotron-4","article":"https://aiwiki.ai/wiki/nemotron_4","kind":"language","developer":"NVIDIA","released":"2024-02-26","parameters":"15B; 340B (Base, Instruct, and Reward checkpoints)","parameters_billions":15,"tier":"partial","weights":"partial","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"The 340B Base, Instruct and Reward checkpoints are downloadable. No public Nemotron-4 15B checkpoint was found.","hugging_face":"https://huggingface.co/nvidia/Nemotron-4-340B-Instruct","article_fact_checked":false},{"name":"Gemma","article":"https://aiwiki.ai/wiki/gemma","kind":"language","developer":"Google DeepMind","released":"2024-02-21","parameters":"270 million to 31 billion across generations","parameters_billions":0.27,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Gemma 4 (April 2026) is Apache 2.0; Gemma 1, 2, 3 and 3n remain under the Gemma Terms of Use, which carry a prohibited use policy.","hugging_face":null,"article_fact_checked":true},{"name":"V-JEPA","article":"https://aiwiki.ai/wiki/v_jepa","kind":"vision","developer":"Meta AI (FAIR)","released":"2024-02-15","parameters":"Up to ~630M (ViT-H/16 encoder)","parameters_billions":0.63,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":false},{"name":"DeepSeekMath","article":"https://aiwiki.ai/wiki/deepseek_math","kind":"language","developer":"DeepSeek","released":"2024-02-05","parameters":"7 billion","parameters_billions":7,"tier":"conditional","weights":"open","license":"deepseek-model","license_name":"DeepSeek Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":false,"note":"Code is MIT; weights use the DeepSeek License. The 120B-token DeepSeekMath-Corpus was not published.","hugging_face":"https://huggingface.co/deepseek-ai/deepseek-math-7b-rl","article_fact_checked":true},{"name":"MiniCPM","article":"https://aiwiki.ai/wiki/minicpm","kind":"language","developer":"OpenBMB","released":"2024-02-01","parameters":"0.43B (MiniCPM4-0.5B) to 13.6B (MiniCPM-MoE-8x2B) across the text models","parameters_billions":0.43,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Apache 2.0 from MiniCPM4 (June 2025). Earlier weights keep OpenBMB licenses that ask for authorization or registration for commercial use.","hugging_face":null,"article_fact_checked":true},{"name":"MiniCPM-V","article":"https://aiwiki.ai/wiki/minicpm_v","kind":"vision","developer":"OpenBMB (THUNLP and ModelBest Inc.)","released":"2024-02-01","parameters":"1.3B to 9B, depending on version","parameters_billions":1.3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Family. Apache 2.0 from v4.0 (latest: 4.6) and for MiniCPM-o 2.6; v1.0 to 2.5 need registration for commercial use; V 2.6 is gated.","hugging_face":null,"article_fact_checked":true},{"name":"OLMo","article":"https://aiwiki.ai/wiki/olmo","kind":"language","developer":"Allen Institute for AI (Ai2)","released":"2024-02-01","parameters":"1B, 7B, 13B, 32B dense; OLMoE 1B active / 7B total","parameters_billions":1,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":true},{"name":"Aya (Expanse / Vision)","article":"https://aiwiki.ai/wiki/aya_expanse","kind":"language","developer":"Cohere Labs (Cohere)","released":"2024-02","parameters":"8B and 32B (Aya Expanse); 13B (Aya 101)","parameters_billions":8,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Aya 23, Aya Expanse and Aya Vision are CC BY-NC 4.0 with Cohere's acceptable use addendum (gated click-through); the earlier Aya 101 is Apache 2.0.","hugging_face":"https://huggingface.co/CohereLabs/aya-expanse-32b","article_fact_checked":false},{"name":"Reka Flash","article":"https://aiwiki.ai/wiki/reka_flash","kind":"language","developer":"Reka AI","released":"2024-02","parameters":"21 billion","parameters_billions":21,"tier":"partial","weights":"partial","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Reka Flash 3 and 3.1 (2025) are Apache 2.0; the original multimodal Reka Flash and its October 2024 update are API only.","hugging_face":"https://huggingface.co/RekaAI/reka-flash-3","article_fact_checked":false},{"name":"AlphaGeometry","article":"https://aiwiki.ai/wiki/alphageometry","kind":"science","developer":"Google DeepMind","released":"2024-01-17","parameters":"Up to 3.3B (AlphaGeometry 2 language model)","parameters_billions":3.3,"tier":"partial","weights":"partial","license":"cc-by-4.0","license_name":"CC BY 4.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"AlphaGeometry 1 checkpoint and vocabulary are CC BY 4.0 (code Apache 2.0); the AlphaGeometry 2 language model weights are not released.","hugging_face":null,"article_fact_checked":true},{"name":"GLM-4","article":"https://aiwiki.ai/wiki/glm_4","kind":"language","developer":"Zhipu AI","released":"2024-01-16","parameters":null,"parameters_billions":null,"tier":"partial","weights":"partial","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"API tiers (GLM-4, Plus, Air, Flash) are closed. GLM-4-0414 weights are MIT; GLM-4-9B uses the glm-4 license (commercial use needs registration).","hugging_face":null,"article_fact_checked":true},{"name":"Minitron","article":"https://aiwiki.ai/wiki/minitron","kind":"language","developer":"NVIDIA","released":"2024","parameters":"About 4B to 8B","parameters_billions":4,"tier":"conditional","weights":"open","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"All base checkpoints carry the NVIDIA Open Model License; the cards for the original Minitron 4B and 8B say research and development only.","hugging_face":null,"article_fact_checked":false},{"name":"InternVL","article":"https://aiwiki.ai/wiki/internvl","kind":"vision","developer":"OpenGVLab (Shanghai AI Laboratory) and academic partners","released":"2023-12-21","parameters":"0.94B to 240.7B (28B active in InternVL3.5-241B-A28B)","parameters_billions":0.94,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"InternVL3.5 is Apache 2.0. Earlier releases are MIT or follow the base LLM (Qwen license on InternVL2.5-78B and InternVL3-78B).","hugging_face":null,"article_fact_checked":true},{"name":"Mixtral","article":"https://aiwiki.ai/wiki/mixtral","kind":"language","developer":"Mistral AI","released":"2023-12-11","parameters":"8x7B: 46.7B total, 12.9B active; 8x22B: 141B total, 39B active","parameters_billions":7,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/mistralai/Mixtral-8x7B-v0.1","article_fact_checked":true},{"name":"Mamba","article":"https://aiwiki.ai/wiki/mamba","kind":"language","developer":"Albert Gu and Tri Dao","released":"2023-12-01","parameters":"130M to 2.8B (pretrained language models)","parameters_billions":0.13,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Architecture page; the license is that of the authors' pretrained language-model checkpoints (130M to 2.8B).","hugging_face":"https://huggingface.co/state-spaces/mamba-2.8b","article_fact_checked":true},{"name":"NVIDIA Parakeet","article":"https://aiwiki.ai/wiki/parakeet","kind":"audio","developer":"NVIDIA (first models with Suno.ai)","released":"2023-12","parameters":"110M to 1.1B","parameters_billions":0.11,"tier":"permissive","weights":"open","license":"cc-by-4.0","license_name":"CC BY 4.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Original models and TDT 0.6B v2 and v3 are CC BY 4.0; newer variants such as parakeet-unified-en-0.6b use the NVIDIA Open Model License.","hugging_face":null,"article_fact_checked":true},{"name":"DeepSeek LLM","article":"https://aiwiki.ai/wiki/deepseek_llm","kind":"language","developer":"DeepSeek","released":"2023-11-29","parameters":"7B and 67B","parameters_billions":7,"tier":"conditional","weights":"open","license":"deepseek-model","license_name":"DeepSeek Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Code is MIT; weights use the DeepSeek License, which allows commercial use subject to use-based restrictions.","hugging_face":"https://huggingface.co/deepseek-ai/deepseek-llm-67b-base","article_fact_checked":false},{"name":"Pika (video generation)","article":"https://aiwiki.ai/wiki/pika","kind":"video","developer":"Pika Labs","released":"2023-11-28","parameters":null,"parameters_billions":null,"tier":"partial","weights":"partial","license":"custom","license_name":"MiniMax H3 Community License Agreement","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Only Pika H3 Flex (LoRA adapters on MiniMax H3, Oct 2026) is open; its license excludes the EU, UK, South Korea and US. Other Pika models are hosted only.","hugging_face":null,"article_fact_checked":true},{"name":"Stable Video Diffusion","article":"https://aiwiki.ai/wiki/stable_video_diffusion","kind":"video","developer":"Stability AI","released":"2023-11-21","parameters":"About 1.52B (1,524,623,082 in the img2vid-xt checkpoint)","parameters_billions":1.52,"tier":"conditional","weights":"open","license":"stability-community","license_name":"Stability AI Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Launched November 2023 as a non-commercial research preview; under the Stability AI Community License since July 2024 (free under $1M revenue).","hugging_face":"https://huggingface.co/stabilityai/stable-video-diffusion-img2vid-xt","article_fact_checked":true},{"name":"GraphCast","article":"https://aiwiki.ai/wiki/graphcast","kind":"science","developer":"Google DeepMind","released":"2023-11-14","parameters":"36.7M","parameters_billions":0.0367,"tier":"permissive","weights":"open","license":"cc-by-4.0","license_name":"CC BY 4.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Weights were relicensed on August 6, 2026 from CC BY-NC-SA 4.0 to CC BY 4.0, which permits commercial use; the code is Apache 2.0.","hugging_face":null,"article_fact_checked":true},{"name":"Grok","article":"https://aiwiki.ai/wiki/grok","kind":"language","developer":"xAI","released":"2023-11-04","parameters":"314B, ~25% active (Grok-1); ~2.7T estimated (Grok-3); later models undisclosed","parameters_billions":314,"tier":"partial","weights":"partial","license":"custom","license_name":"Grok 2 Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Grok-1 is Apache 2.0; Grok 2 weights (Aug 2025) allow commercial use under xAI's usage policy; Grok 3 and later are closed.","hugging_face":null,"article_fact_checked":true},{"name":"DeepSeek-Coder","article":"https://aiwiki.ai/wiki/deepseek_coder","kind":"language","developer":"DeepSeek","released":"2023-11-02","parameters":"1.3B, 6.7B, 33B dense (V1); 16B and 236B MoE with 2.4B and 21B active (V2)","parameters_billions":1.3,"tier":"conditional","weights":"open","license":"deepseek-model","license_name":"DeepSeek Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Code is MIT; V1 and V2 weights use the DeepSeek License, which allows commercial use subject to use-based restrictions.","hugging_face":"https://huggingface.co/deepseek-ai/deepseek-coder-33b-instruct","article_fact_checked":true},{"name":"Yi (language model)","article":"https://aiwiki.ai/wiki/yi","kind":"language","developer":"01.AI","released":"2023-11-02","parameters":"6B and 34B base (family spans 1.5B to 34B)","parameters_billions":6,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Yi-6B and Yi-34B launched under the Yi Series Models Community License; the whole open Yi series is now Apache 2.0.","hugging_face":"https://huggingface.co/01-ai/Yi-34B","article_fact_checked":false},{"name":"Nemotron","article":"https://aiwiki.ai/wiki/nemotron","kind":"language","developer":"NVIDIA","released":"2023-11","parameters":"8B (2023) to ~550B total MoE (Nemotron 3 Ultra, 2026)","parameters_billions":8,"tier":"permissive","weights":"open","license":"openmdw-1.1","license_name":"OpenMDW 1.1","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Nemotron 3 Ultra and the 3.5 line use OpenMDW 1.1; earlier models use NVIDIA open model licenses; Nemotron-H is research only.","hugging_face":null,"article_fact_checked":true},{"name":"BitNet","article":"https://aiwiki.ai/wiki/bitnet","kind":"language","developer":"Microsoft Research","released":"2023-10-17","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/microsoft/bitnet-b1.58-2B-4T","article_fact_checked":false},{"name":"Latent Consistency Models (LCM)","article":"https://aiwiki.ai/wiki/latent_consistency_models","kind":"image","developer":"Tsinghua University (IIIS)","released":"2023-10-06","parameters":null,"parameters_billions":null,"tier":"conditional","weights":"open","license":"openrail","license_name":"OpenRAIL","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Official LCM-LoRA and LCM SDXL/SSD-1B checkpoints are OpenRAIL++-M; the first LCM_Dreamshaper_v7 checkpoint is tagged MIT.","hugging_face":null,"article_fact_checked":false},{"name":"CogVLM","article":"https://aiwiki.ai/wiki/cogvlm","kind":"vision","developer":"Zhipu AI and Tsinghua University (KEG)","released":"2023-10-05","parameters":"About 17B (about 10B vision, 7B language)","parameters_billions":17,"tier":"conditional","weights":"open","license":"custom","license_name":"CogVLM License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Free for academic research; commercial use requires registration. CogVLM2 adds the Llama 3 license terms. Code is Apache 2.0.","hugging_face":null,"article_fact_checked":false},{"name":"SEDD (Score Entropy Discrete Diffusion)","article":"https://aiwiki.ai/wiki/sedd","kind":"language","developer":"Stanford University (Aaron Lou, Chenlin Meng, Stefano Ermon)","released":"2023-10","parameters":null,"parameters_billions":null,"tier":"conditional","weights":"open","license":"unspecified","license_name":"License not stated","commercial_use":null,"training_code_published":true,"training_data_published":null,"note":"The code repository is MIT and links the checkpoints; the Hugging Face checkpoint repos carry no license statement of their own.","hugging_face":null,"article_fact_checked":false},{"name":"XTTS (Coqui XTTS)","article":"https://aiwiki.ai/wiki/xtts","kind":"audio","developer":"Coqui AI","released":"2023-09-30","parameters":"About 750M","parameters_billions":0.75,"tier":"noncommercial","weights":"open","license":"custom","license_name":"Coqui Public Model License 1.0.0","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Coqui AI shut down in January 2024, so the separate commercial license it used to sell is no longer available.","hugging_face":"https://huggingface.co/coqui/XTTS-v2","article_fact_checked":false},{"name":"Mistral 7B","article":"https://aiwiki.ai/wiki/mistral_7b","kind":"language","developer":"Mistral AI","released":"2023-09-27","parameters":"~7.24 billion (announced as 7.3B)","parameters_billions":7.24,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/mistralai/Mistral-7B-v0.1","article_fact_checked":true},{"name":"Hunyuan","article":"https://aiwiki.ai/wiki/hunyuan","kind":"language","developer":"Tencent","released":"2023-09-07","parameters":"389B total, 52B active (Hunyuan-Large); 560B total, 56B active (TurboS)","parameters_billions":389,"tier":"partial","weights":"partial","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Hy3 and Hy4 Preview are Apache 2.0; earlier open releases use Tencent community licenses that exclude the EU, UK and South Korea; T1 and TurboS are hosted only.","hugging_face":null,"article_fact_checked":true},{"name":"Nougat (model)","article":"https://aiwiki.ai/wiki/nougat","kind":"vision","developer":"Meta AI","released":"2023-08-25","parameters":"250M (small); 350M (base)","parameters_billions":0.25,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Code is MIT. The nougat-small repository is tagged CC BY 4.0 on Hugging Face, but Meta's README says the weights are CC-BY-NC.","hugging_face":"https://huggingface.co/facebook/nougat-base","article_fact_checked":false},{"name":"Code Llama","article":"https://aiwiki.ai/wiki/code_llama","kind":"language","developer":"Meta AI","released":"2023-08-24","parameters":"7B, 13B, 34B, 70B (base, Python, and Instruct variants)","parameters_billions":7,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Llama 2 Community License: acceptable use policy applies and a separate license is needed above 700M monthly users.","hugging_face":"https://huggingface.co/codellama/CodeLlama-70b-hf","article_fact_checked":true},{"name":"Qwen-VL","article":"https://aiwiki.ai/wiki/qwen_vl","kind":"vision","developer":"Alibaba Cloud (Qwen team)","released":"2023-08-22","parameters":"9.6B","parameters_billions":9.6,"tier":"partial","weights":"partial","license":"qwen-license","license_name":"Qwen License (Tongyi Qianwen)","commercial_use":"conditional","training_code_published":true,"training_data_published":null,"note":"Qwen-VL and Qwen-VL-Chat (9.6B) are open; Qwen-VL-Plus and Qwen-VL-Max are API only. Above 100M monthly users a separate license is needed.","hugging_face":null,"article_fact_checked":true},{"name":"SeamlessM4T","article":"https://aiwiki.ai/wiki/seamless_m4t","kind":"audio","developer":"Meta AI","released":"2023-08-22","parameters":"1.2B (Medium), 2.3B (Large and v2)","parameters_billions":1.2,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Supporting code and the w2v-BERT 2.0 encoder are MIT; SeamlessExpressive uses a separate Seamless non-commercial license.","hugging_face":"https://huggingface.co/facebook/seamless-m4t-v2-large","article_fact_checked":false},{"name":"AudioCraft","article":"https://aiwiki.ai/wiki/audiocraft","kind":"audio","developer":"Meta AI (FAIR)","released":"2023-08-02","parameters":"300M to 3.3B (MusicGen); 1.5B (AudioGen)","parameters_billions":0.3,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Code and training framework are MIT; the MusicGen, AudioGen and MAGNeT weights are CC BY-NC 4.0.","hugging_face":null,"article_fact_checked":false},{"name":"BGE (BAAI General Embedding)","article":"https://aiwiki.ai/wiki/bge","kind":"other","developer":"Beijing Academy of Artificial Intelligence (BAAI)","released":"2023-08-02","parameters":"24M to 326M (v1.5 embedders)","parameters_billions":0.024,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"v1.5, BGE-M3 and BGE-VL are MIT; bge-en-icl, bge-reranker-v2-m3 and bge-reasoner are Apache 2.0; the Gemma-2-based models carry Gemma terms.","hugging_face":null,"article_fact_checked":false},{"name":"SDXL (Stable Diffusion XL)","article":"https://aiwiki.ai/wiki/sdxl","kind":"image","developer":"Stability AI","released":"2023-07-26","parameters":"~2.6B U-Net; ~6.6B total with refiner","parameters_billions":2.6,"tier":"conditional","weights":"open","license":"openrail","license_name":"OpenRAIL","commercial_use":"conditional","training_code_published":null,"training_data_published":false,"note":"SDXL 1.0 base and refiner are CreativeML Open RAIL++-M; SDXL Turbo is under a non-commercial Stability AI license.","hugging_face":"https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0","article_fact_checked":true},{"name":"Llama 2","article":"https://aiwiki.ai/wiki/llama_2","kind":"language","developer":"Meta","released":"2023-07-18","parameters":"7B, 13B, 70B (34B trained but withheld)","parameters_billions":7,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Gated download. Over 700M monthly active users needs a separate Meta license; outputs may not be used to improve other LLMs.","hugging_face":null,"article_fact_checked":true},{"name":"InternLM","article":"https://aiwiki.ai/wiki/internlm","kind":"language","developer":"Shanghai AI Laboratory","released":"2023-07-06","parameters":"1.8B to 20B across generations; single 8B flagship in InternLM3","parameters_billions":1.8,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"InternLM3 and Intern-S1 are Apache 2.0; InternLM, InternLM2 and InternLM2.5 weights need an application form for commercial use.","hugging_face":"https://huggingface.co/internlm/internlm3-8b-instruct","article_fact_checked":true},{"name":"MusicGen","article":"https://aiwiki.ai/wiki/musicgen","kind":"audio","developer":"Meta AI (FAIR)","released":"2023-06-08","parameters":"300M (Small), 1.5B (Medium), 3.3B (Large)","parameters_billions":0.3,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Weights are CC BY-NC 4.0; the AudioCraft code, including training code, is MIT.","hugging_face":"https://huggingface.co/facebook/musicgen-large","article_fact_checked":true},{"name":"Hiera","article":"https://aiwiki.ai/wiki/hiera","kind":"vision","developer":"Meta AI (FAIR)","released":"2023-06-01","parameters":"28M to 673M","parameters_billions":0.028,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Model weights are CC BY-NC 4.0; the code was relicensed to Apache 2.0 in March 2024.","hugging_face":null,"article_fact_checked":false},{"name":"Würstchen","article":"https://aiwiki.ai/wiki/wuerstchen","kind":"image","developer":"Pablo Pernias, Dominic Rampas, Mats L. Richter, Christopher J. Pal, Marc Aubreville","released":"2023-06-01","parameters":"Stage A 18M; Stage B ~1B; Stage C ~0.99B","parameters_billions":0.018000000000000002,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/warp-ai/wuerstchen","article_fact_checked":false},{"name":"Aquila (language model)","article":"https://aiwiki.ai/wiki/aquila","kind":"language","developer":"Beijing Academy of Artificial Intelligence (BAAI)","released":"2023-06","parameters":"7B, 34B, and experimental 70B","parameters_billions":7,"tier":"conditional","weights":"open","license":"custom","license_name":"BAAI Aquila Model License Agreement","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Commercial use is allowed within the license restrictions; the experimental 70B models use a separate BAAI Aquila 70B license. Code is Apache 2.0.","hugging_face":null,"article_fact_checked":false},{"name":"Phi (language model)","article":"https://aiwiki.ai/wiki/phi","kind":"language","developer":"Microsoft Research","released":"2023-06","parameters":"1.3B (Phi-1) to 15B (Phi-4-reasoning-vision)","parameters_billions":1.3,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/microsoft/phi-4","article_fact_checked":true},{"name":"Massively Multilingual Speech (MMS)","article":"https://aiwiki.ai/wiki/massively_multilingual_speech","kind":"audio","developer":"Meta AI","released":"2023-05-22","parameters":"317M (MMS-300M), 965M (MMS-1B)","parameters_billions":0.317,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":false},{"name":"ImageBind","article":"https://aiwiki.ai/wiki/imagebind","kind":"vision","developer":"Meta AI (FAIR)","released":"2023-05-09","parameters":null,"parameters_billions":null,"tier":"noncommercial","weights":"open","license":"cc-by-nc-sa-4.0","license_name":"CC BY-NC-SA 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"The README says CC-BY-NC 4.0, but the LICENSE file in the repository is CC BY-NC-SA 4.0. Weights are downloaded from GitHub.","hugging_face":null,"article_fact_checked":true},{"name":"StarCoder","article":"https://aiwiki.ai/wiki/starcoder","kind":"language","developer":"BigCode project (Hugging Face and ServiceNow)","released":"2023-05-04","parameters":"15.5B (StarCoder); 3B, 7B, 15B (StarCoder2)","parameters_billions":15.5,"tier":"conditional","weights":"open","license":"openrail","license_name":"OpenRAIL","commercial_use":"conditional","training_code_published":null,"training_data_published":true,"note":"BigCode OpenRAIL-M with use restrictions for StarCoder and StarCoder2; the Stack training datasets are published on Hugging Face.","hugging_face":"https://huggingface.co/bigcode/starcoder","article_fact_checked":true},{"name":"Shap-E","article":"https://aiwiki.ai/wiki/shap_e","kind":"other","developer":"OpenAI","released":"2023-05-03","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/openai/shap-e","article_fact_checked":false},{"name":"Falcon (language model)","article":"https://aiwiki.ai/wiki/falcon","kind":"language","developer":"Technology Innovation Institute (TII)","released":"2023-05","parameters":"90M to 180B across generations","parameters_billions":0.09,"tier":"conditional","weights":"open","license":"falcon-llm","license_name":"TII Falcon License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Falcon 3 and Falcon-H1 use the TII Falcon LLM License; Falcon 7B and 40B are Apache 2.0; Falcon 180B has its own stricter TII license.","hugging_face":null,"article_fact_checked":true},{"name":"WizardLM","article":"https://aiwiki.ai/wiki/wizardlm","kind":"language","developer":"Microsoft","released":"2023-04-24","parameters":null,"parameters_billions":null,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":true,"note":"License follows each base model; the latest hosted chat models are Llama 2 Community. WizardLM-2 was withdrawn from official channels.","hugging_face":null,"article_fact_checked":false},{"name":"DINOv2","article":"https://aiwiki.ai/wiki/dinov2","kind":"vision","developer":"Meta AI Research (FAIR)","released":"2023-04-14","parameters":"21M to 1.1B (ViT-S/14, ViT-B/14, ViT-L/14, ViT-g/14)","parameters_billions":0.021,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":false,"note":"Relicensed from CC BY-NC 4.0 to Apache 2.0 on August 31, 2023. The LVD-142M training set was not released.","hugging_face":"https://huggingface.co/facebook/dinov2-giant","article_fact_checked":true},{"name":"Qwen","article":"https://aiwiki.ai/wiki/qwen","kind":"language","developer":"Alibaba Cloud (Qwen Team)","released":"2023-04-11","parameters":"Roughly 0.5B to over 1T across variants","parameters_billions":0.5,"tier":"partial","weights":"partial","license":"custom","license_name":"Qwen Community License 1.0","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Latest open release (Qwen3.8-Flash-Next). Most Qwen3 to Qwen3.8 checkpoints are Apache 2.0; Qwen3.8-Max has its own license; some tiers are hosted only.","hugging_face":null,"article_fact_checked":true},{"name":"SenseNova","article":"https://aiwiki.ai/wiki/sensenova","kind":"language","developer":"SenseTime","released":"2023-04-10","parameters":"2B to 27B (open-weight models); over 600B (SenseNova V6)","parameters_billions":2,"tier":"partial","weights":"partial","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Flagship generations are API only. Open releases (U1, U1.5, Code2Video) are mostly Apache 2.0; some are MIT and SenseNova-Vision is CC BY-NC 4.0.","hugging_face":null,"article_fact_checked":true},{"name":"Segment Anything Model and Dataset (SAM and SA-1B)","article":"https://aiwiki.ai/wiki/segment_anything_model_and_dataset_sam_and_sa-1b","kind":"vision","developer":"Meta AI","released":"2023-04-05","parameters":"~91M (ViT-B), ~308M (ViT-L), ~636M (ViT-H)","parameters_billions":0.091,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":true,"note":"SAM weights and code are Apache 2.0; the SA-1B training dataset is downloadable but under a separate research-only license.","hugging_face":"https://huggingface.co/facebook/sam-vit-huge","article_fact_checked":true},{"name":"DINO (computer vision)","article":"https://aiwiki.ai/wiki/dino_model","kind":"vision","developer":"Meta AI","released":"2023-04","parameters":"21M (ViT-S) to 6.7B (DINOv3 ViT-7B)","parameters_billions":0.021,"tier":"conditional","weights":"open","license":"custom","license_name":"DINOv3 License","commercial_use":"conditional","training_code_published":true,"training_data_published":null,"note":"DINO and DINOv2 are Apache 2.0. DINOv3 allows commercial use but is gated and bars military, nuclear and other trade-controlled end uses.","hugging_face":null,"article_fact_checked":false},{"name":"LLaVA (Large Language and Vision Assistant)","article":"https://aiwiki.ai/wiki/llava","kind":"vision","developer":"University of Wisconsin-Madison, Microsoft Research and Columbia University","released":"2023-04","parameters":"7B and 13B backbones (LLaVA-1.5); 0.5B to 110B backbones in later versions","parameters_billions":7,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Family. OneVision and OneVision-1.5 (latest) are Apache 2.0; Vicuna-based 1.5 and 1.6 checkpoints cite the Llama 2 Community License.","hugging_face":null,"article_fact_checked":true},{"name":"RFdiffusion","article":"https://aiwiki.ai/wiki/rfdiffusion","kind":"science","developer":"Baker Lab, UW Institute for Protein Design","released":"2023-03-30","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"bsd-3-clause","license_name":"BSD 3-Clause","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":false},{"name":"Vicuna (language model)","article":"https://aiwiki.ai/wiki/vicuna","kind":"language","developer":"LMSYS","released":"2023-03-30","parameters":"7B, 13B, 33B","parameters_billions":7,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":true,"training_data_published":false,"note":"v1.5 (latest) is under the Llama 2 Community License; v0 to v1.3, including the 33B model, inherit the non-commercial LLaMA license.","hugging_face":"https://huggingface.co/lmsys/vicuna-13b-v1.5","article_fact_checked":true},{"name":"SigLIP","article":"https://aiwiki.ai/wiki/siglip","kind":"vision","developer":"Google DeepMind","released":"2023-03-27","parameters":"86M to 1B (SigLIP 2, ViT-B to ViT-g)","parameters_billions":0.08600000000000001,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":false,"note":"SigLIP and SigLIP 2 checkpoints are Apache 2.0; the WebLI training set is not public.","hugging_face":null,"article_fact_checked":true},{"name":"ChatGLM","article":"https://aiwiki.ai/wiki/chatglm","kind":"language","developer":"Zhipu AI with Tsinghua University KEG lab","released":"2023-03-14","parameters":"~6.2B (ChatGLM-6B); ~6B per later generation","parameters_billions":6.2,"tier":"conditional","weights":"open","license":"custom","license_name":"ChatGLM Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Each generation (ChatGLM-6B, ChatGLM2-6B, ChatGLM3-6B) has its own Model License: free for research, commercial use after registering with Zhipu AI.","hugging_face":"https://huggingface.co/zai-org/chatglm-6b","article_fact_checked":true},{"name":"Alpaca (model)","article":"https://aiwiki.ai/wiki/alpaca","kind":"other","developer":"Stanford CRFM","released":"2023-03-13","parameters":"7B (13B variant documented)","parameters_billions":7,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":true,"note":"Released as a weight diff; rebuilding the model needs Meta's original LLaMA 7B weights. Research use only.","hugging_face":null,"article_fact_checked":true},{"name":"CodeGeeX","article":"https://aiwiki.ai/wiki/codegeex","kind":"language","developer":"Tsinghua University (KEG) and Zhipu AI","released":"2023-03","parameters":"6B to 13B","parameters_billions":6,"tier":"conditional","weights":"open","license":"custom","license_name":"CodeGeeX4 License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Latest release (CodeGeeX4-ALL-9B): free for academic research, commercial use needs registration. The original 13B CodeGeeX License is non-commercial only.","hugging_face":null,"article_fact_checked":false},{"name":"LLaMA","article":"https://aiwiki.ai/wiki/llama","kind":"language","developer":"Meta AI","released":"2023-02-24","parameters":"Up to 65B (LLaMA 1); ~2T total (Llama 4 Behemoth, unreleased)","parameters_billions":65,"tier":"conditional","weights":"open","license":"llama-community","license_name":"Llama Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"LLaMA 1 was research-only on request; Llama 2 to 4 use versioned Community Licenses; Llama 4 Behemoth was never released.","hugging_face":null,"article_fact_checked":true},{"name":"ControlNet","article":"https://aiwiki.ai/wiki/controlnet","kind":"image","developer":"Lvmin Zhang, Anyi Rao and Maneesh Agrawala (Stanford University)","released":"2023-02-10","parameters":null,"parameters_billions":null,"tier":"conditional","weights":"open","license":"openrail","license_name":"OpenRAIL","commercial_use":"conditional","training_code_published":true,"training_data_published":null,"note":"Official Stable Diffusion 1.5 checkpoints (1.0 and 1.1) are tagged OpenRAIL; the code is Apache 2.0. Third-party SDXL ControlNets carry their own licenses.","hugging_face":null,"article_fact_checked":true},{"name":"YOLOv8","article":"https://aiwiki.ai/wiki/yolov8","kind":"vision","developer":"Ultralytics","released":"2023-01-10","parameters":"3.2M (YOLOv8n) to 68.2M (YOLOv8x) for detection","parameters_billions":0.0032,"tier":"conditional","weights":"open","license":"custom","license_name":"GNU AGPL-3.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Copyleft: GPL-3.0 at launch, AGPL-3.0 since April 14, 2023. Ultralytics also sells an Enterprise license.","hugging_face":"https://huggingface.co/Ultralytics/YOLOv8","article_fact_checked":false},{"name":"SantaCoder","article":"https://aiwiki.ai/wiki/santacoder","kind":"language","developer":"BigCode (Hugging Face and ServiceNow Research)","released":"2023-01-09","parameters":"1.1B","parameters_billions":1.1,"tier":"conditional","weights":"open","license":"openrail","license_name":"OpenRAIL","commercial_use":"conditional","training_code_published":null,"training_data_published":true,"note":"BigCode OpenRAIL-M: commercial use allowed subject to its use restrictions. Trained on The Stack v1.1, which BigCode publishes.","hugging_face":"https://huggingface.co/bigcode/santacoder","article_fact_checked":false},{"name":"ConvNeXt","article":"https://aiwiki.ai/wiki/convnext","kind":"vision","developer":"Facebook AI Research (FAIR) and UC Berkeley","released":"2023-01","parameters":"3.7M (V2-Atto) to 660M (V2-Huge); original ConvNeXt 29M to 350M","parameters_billions":0.0037,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Original ConvNeXt (2022) checkpoints are MIT, commercial use allowed. ConvNeXt V2 (2023) pretrained and fine-tuned checkpoints are CC BY-NC 4.0.","hugging_face":null,"article_fact_checked":true},{"name":"I-JEPA","article":"https://aiwiki.ai/wiki/i_jepa","kind":"vision","developer":"Meta AI","released":"2023-01","parameters":"About 632M (ViT-Huge)","parameters_billions":0.632,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Code and pretrained checkpoints share the repository's CC BY-NC 4.0 license.","hugging_face":null,"article_fact_checked":false},{"name":"MetaCLIP","article":"https://aiwiki.ai/wiki/metaclip","kind":"vision","developer":"Meta AI","released":"2023","parameters":null,"parameters_billions":null,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"MetaCLIP and MetaCLIP 2 checkpoints are CC BY-NC 4.0; the bundled open_clip code keeps its own license.","hugging_face":null,"article_fact_checked":false},{"name":"InternVideo","article":"https://aiwiki.ai/wiki/internvideo","kind":"vision","developer":"OpenGVLab, Shanghai AI Laboratory; Nanjing University; SIAT, CAS","released":"2022-12-06","parameters":"Up to 6B (InternVideo2-6B)","parameters_billions":6,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Most checkpoints are Apache 2.0; a few (ViCLIP, InternVideo2-Chat-8B) are MIT. Some sit behind a Hugging Face click-through gate.","hugging_face":null,"article_fact_checked":false},{"name":"Galactica (language model)","article":"https://aiwiki.ai/wiki/galactica","kind":"science","developer":"Meta AI (Papers with Code team)","released":"2022-11-15","parameters":"Five sizes, 125M to 120B","parameters_billions":0.125,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"All five sizes are on Hugging Face under CC BY-NC 4.0; the galai inference library is Apache 2.0.","hugging_face":"https://huggingface.co/facebook/galactica-120b","article_fact_checked":false},{"name":"ESMFold","article":"https://aiwiki.ai/wiki/esmfold","kind":"science","developer":"Meta AI (FAIR Protein Team)","released":"2022-11-01","parameters":"Production checkpoints use 3B ESM-2 stem; ESM-2 backbone available at 8M to 15B","parameters_billions":3,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/facebook/esmfold_v1","article_fact_checked":true},{"name":"EnCodec","article":"https://aiwiki.ai/wiki/encodec","kind":"audio","developer":"Meta AI (FAIR)","released":"2022-10-24","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Reference 24 and 48 kHz checkpoints ship with the MIT repository; the 32 kHz AudioCraft checkpoint used by MusicGen is CC BY-NC 4.0.","hugging_face":null,"article_fact_checked":true},{"name":"Runwayml/stable-diffusion-v1-5 model","article":"https://aiwiki.ai/wiki/runwayml_stable-diffusion-v1-5_model","kind":"image","developer":"Runway (RunwayML)","released":"2022-10-20","parameters":"~860M U-Net (trainable); ~123M CLIP text encoder; ~84M VAE","parameters_billions":0.86,"tier":"conditional","weights":"open","license":"openrail","license_name":"OpenRAIL","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Runway deleted its original repository in August 2024; hash-identical community mirrors keep the CreativeML Open RAIL-M license.","hugging_face":"https://huggingface.co/stable-diffusion-v1-5/stable-diffusion-v1-5","article_fact_checked":false},{"name":"BioGPT","article":"https://aiwiki.ai/wiki/biogpt","kind":"science","developer":"Microsoft Research","released":"2022-10","parameters":"347M (base); 1.5B (BioGPT-Large)","parameters_billions":0.34700000000000003,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/microsoft/biogpt","article_fact_checked":false},{"name":"Whisper","article":"https://aiwiki.ai/wiki/whisper","kind":"audio","developer":"OpenAI","released":"2022-09-21","parameters":"39M (tiny) to 1,550M (large)","parameters_billions":0.039,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"OpenAI's GitHub repository states MIT for code and weights; some Hugging Face cards (large-v3 among them) are tagged Apache 2.0.","hugging_face":"https://huggingface.co/openai/whisper-large-v3","article_fact_checked":true},{"name":"Stable Diffusion","article":"https://aiwiki.ai/wiki/stable_diffusion","kind":"image","developer":"Stability AI, with CompVis (LMU Munich), Runway, LAION, and EleutherAI","released":"2022-08-22","parameters":"860M U-Net (SD 1.x); ~3.5B (SDXL base); 8.1B (SD 3.5 Large)","parameters_billions":0.86,"tier":"conditional","weights":"open","license":"stability-community","license_name":"Stability AI Community License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"SD3 and SD 3.5 use the Stability AI Community License (free under $1M annual revenue); v1, v2 and SDXL use OpenRAIL licenses.","hugging_face":null,"article_fact_checked":true},{"name":"Atlas (language model)","article":"https://aiwiki.ai/wiki/atlas_language_model","kind":"language","developer":"Meta AI (FAIR)","released":"2022-08","parameters":"Reader 220M to 11B, plus ~110M retriever","parameters_billions":0.22,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"The repository license is CC BY-NC 4.0; no separate license is stated for the checkpoints. Wikipedia-derived corpora and indices are CC BY-SA.","hugging_face":null,"article_fact_checked":false},{"name":"GLM-130B","article":"https://aiwiki.ai/wiki/glm_130b","kind":"language","developer":"Tsinghua KEG and Zhipu AI","released":"2022-08","parameters":"130 billion (dense)","parameters_billions":130,"tier":"noncommercial","weights":"open","license":"custom","license_name":"GLM-130B License","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Weights are sent after an application form; the license allows non-commercial research only. Code is Apache 2.0.","hugging_face":null,"article_fact_checked":false},{"name":"BLOOM (language model)","article":"https://aiwiki.ai/wiki/bloom","kind":"language","developer":"BigScience project (coordinated by Hugging Face)","released":"2022-07-12","parameters":"176 billion (smaller variants 560M to 7.1B)","parameters_billions":176,"tier":"conditional","weights":"open","license":"openrail","license_name":"OpenRAIL","commercial_use":"conditional","training_code_published":null,"training_data_published":false,"note":"BigScience RAIL License v1.0: commercial use allowed subject to its use restrictions. The ROOTS training corpus was not released in full.","hugging_face":"https://huggingface.co/bigscience/bloom","article_fact_checked":true},{"name":"No Language Left Behind (NLLB)","article":"https://aiwiki.ai/wiki/no_language_left_behind","kind":"language","developer":"Meta AI","released":"2022-07","parameters":"600M to 54.5B","parameters_billions":0.6,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"All NLLB-200 checkpoints are CC BY-NC 4.0; the fairseq code is MIT.","hugging_face":null,"article_fact_checked":false},{"name":"OpenFold","article":"https://aiwiki.ai/wiki/openfold","kind":"science","developer":"AlQuraishi lab (Columbia University) and the OpenFold Consortium","released":"2022-06","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"cc-by-4.0","license_name":"CC BY 4.0","commercial_use":"yes","training_code_published":true,"training_data_published":true,"note":"Retrained OpenFold weights and OpenProteinSet are CC BY 4.0; code is Apache 2.0. The successor OpenFold3 is Apache 2.0.","hugging_face":null,"article_fact_checked":true},{"name":"OPT (Open Pre-trained Transformer)","article":"https://aiwiki.ai/wiki/opt","kind":"language","developer":"Meta AI","released":"2022-05-03","parameters":"125M to 175B (9 sizes)","parameters_billions":0.125,"tier":"noncommercial","weights":"open","license":"custom","license_name":"OPT-175B License Agreement","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Sizes 125M to 66B download directly; OPT-175B weights are granted on request to researchers. Non-commercial research use only.","hugging_face":null,"article_fact_checked":false},{"name":"GLIDE (OpenAI)","article":"https://aiwiki.ai/wiki/glide","kind":"image","developer":"OpenAI","released":"2021-12-20","parameters":"3.5B (full model); ~300M released filtered model","parameters_billions":3.5,"tier":"partial","weights":"partial","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Only the small GLIDE (filtered) model (~300M) was released; the full 3.5B model was withheld. OpenAI's model card does not recommend commercial use.","hugging_face":null,"article_fact_checked":false},{"name":"EXAONE","article":"https://aiwiki.ai/wiki/exaone","kind":"language","developer":"LG AI Research (LG Corporation)","released":"2021-12-14","parameters":"1.2B to 32B dense; 236B-A23B (K-EXAONE) and 750B-A37B (K-EXAONE 2.0) MoE","parameters_billions":1.2,"tier":"partial","weights":"partial","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"K-EXAONE 2.0 (Jul 2026) is Apache 2.0; EXAONE 3.0 to 4.5 use a non-commercial license; K-EXAONE 236B has its own license; 1.0 and 2.0 were closed.","hugging_face":null,"article_fact_checked":false},{"name":"Donut (Model)","article":"https://aiwiki.ai/wiki/donut_model","kind":"vision","developer":"NAVER CLOVA","released":"2021-11","parameters":"143M (176M when English token embeddings are counted)","parameters_billions":0.14300000000000002,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/naver-clova-ix/donut-base","article_fact_checked":true},{"name":"AlphaFold-Multimer","article":"https://aiwiki.ai/wiki/alphafold_multimer","kind":"science","developer":"Google DeepMind","released":"2021-10-04","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"cc-by-4.0","license_name":"CC BY 4.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Model parameters are CC BY 4.0; the code is Apache 2.0. Multimer v1 and v2 weights remain available but are deprecated.","hugging_face":null,"article_fact_checked":false},{"name":"RWKV","article":"https://aiwiki.ai/wiki/rwkv","kind":"other","developer":"Bo Peng (BlinkDL) and the RWKV project, hosted by the Linux Foundation","released":"2021-08","parameters":"169M to 14B","parameters_billions":0.169,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Architecture and model family; the official RWKV-4 to RWKV-7 checkpoints are all Apache 2.0.","hugging_face":null,"article_fact_checked":true},{"name":"Sentence-transformers/all-MiniLM-L6-v2 model","article":"https://aiwiki.ai/wiki/sentence-transformers_all-minilm-l6-v2_model","kind":"language","developer":"Sentence Transformers project (2021 Hugging Face JAX/Flax community sprint)","released":"2021-08","parameters":"~22.7 million","parameters_billions":0.0227,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2","article_fact_checked":false},{"name":"Sentence-transformers/all-mpnet-base-v2 model","article":"https://aiwiki.ai/wiki/sentence-transformers_all-mpnet-base-v2_model","kind":"language","developer":"Sentence Transformers project (2021 Hugging Face Flax/JAX community sprint)","released":"2021-08","parameters":"~110 million","parameters_billions":0.11,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/sentence-transformers/all-mpnet-base-v2","article_fact_checked":false},{"name":"AlphaFold","article":"https://aiwiki.ai/wiki/alphafold","kind":"other","developer":"Google DeepMind","released":"2021-07-15","parameters":null,"parameters_billions":null,"tier":"noncommercial","weights":"open","license":"custom","license_name":"AlphaFold 3 Model Parameters Terms of Use","commercial_use":"no","training_code_published":null,"training_data_published":null,"note":"Latest release (AlphaFold 3) is non-commercial only. AlphaFold 2 and AlphaFold-Multimer parameters are CC BY 4.0 (commercial use allowed). Code is Apache 2.0.","hugging_face":null,"article_fact_checked":true},{"name":"SPLADE","article":"https://aiwiki.ai/wiki/splade","kind":"language","developer":"Naver Labs Europe","released":"2021-07-12","parameters":null,"parameters_billions":null,"tier":"noncommercial","weights":"open","license":"cc-by-nc-sa-4.0","license_name":"CC BY-NC-SA 4.0","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Code and checkpoints from SPLADE v2 to SPLADE-v3 are CC BY-NC-SA 4.0; naver/splade-v3 is gated on Hugging Face.","hugging_face":null,"article_fact_checked":true},{"name":"HuBERT","article":"https://aiwiki.ai/wiki/hubert","kind":"audio","developer":"Meta AI (then Facebook AI Research)","released":"2021-06-14","parameters":"About 95M (Base), 317M (Large), 964M (X-Large)","parameters_billions":0.095,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/facebook/hubert-large-ls960-ft","article_fact_checked":true},{"name":"GPT-J","article":"https://aiwiki.ai/wiki/gpt_j","kind":"language","developer":"EleutherAI","released":"2021-06-09","parameters":"6,053,381,344 (~6.05B)","parameters_billions":6.05,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/EleutherAI/gpt-j-6b","article_fact_checked":true},{"name":"GLM","article":"https://aiwiki.ai/wiki/glm","kind":"language","developer":"Tsinghua University and Zhipu AI (Z.ai)","released":"2021-06","parameters":"110M to 744B total (40B active, GLM-5)","parameters_billions":0.11,"tier":"partial","weights":"partial","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"GLM-4.5 to 5.2 and GLM-5.3-Flash are MIT; the flagship GLM-5.3 has a custom license; GLM-4 mixed API-only and open 9B models.","hugging_face":null,"article_fact_checked":false},{"name":"Huawei PanGu","article":"https://aiwiki.ai/wiki/pangu","kind":"language","developer":"Huawei (Huawei Cloud and Noah's Ark Lab)","released":"2021-04","parameters":"1.3B to 1.085T across the family","parameters_billions":1.3,"tier":"partial","weights":"partial","license":"custom","license_name":"openPangu Model License Agreement 1.0","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Only the 2025 openPangu releases (Embedded-7B, Pangu Pro MoE 72B) are open. The license bars use in the EU and requires a Powered by openPangu notice.","hugging_face":null,"article_fact_checked":false},{"name":"Switch Transformer","article":"https://aiwiki.ai/wiki/switch_transformer","kind":"language","developer":"Google (Google Brain)","released":"2021-01-11","parameters":"Up to 1.571 trillion (Switch-C, 2,048 experts)","parameters_billions":1571,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":true},{"name":"CLIP (Contrastive Language-Image Pre-training)","article":"https://aiwiki.ai/wiki/clip","kind":"vision","developer":"OpenAI","released":"2021","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":false,"note":"MIT is the license of OpenAI's repository, which distributes the nine checkpoints; no separate weights license is stated. Training data is not released.","hugging_face":null,"article_fact_checked":true},{"name":"Swin Transformer","article":"https://aiwiki.ai/wiki/swin_transformer","kind":"vision","developer":"Microsoft Research Asia","released":"2021","parameters":"29M (Swin-T) to 197M (Swin-L); about 3B (SwinV2-G)","parameters_billions":0.029,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Architecture page; the official repository with its pretrained models is MIT. Hugging Face ports (microsoft/swin-*) are tagged Apache 2.0.","hugging_face":null,"article_fact_checked":true},{"name":"DeiT","article":"https://aiwiki.ai/wiki/deit","kind":"vision","developer":"Facebook AI Research (FAIR) and Sorbonne University","released":"2020-12-23","parameters":"5M to 87M","parameters_billions":0.005,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"The facebookresearch/deit repository (DeiT and DeiT III code and checkpoints) is Apache 2.0.","hugging_face":null,"article_fact_checked":false},{"name":"mT5","article":"https://aiwiki.ai/wiki/mt5","kind":"language","developer":"Google Research","released":"2020-10-22","parameters":"300M to 13B (five sizes)","parameters_billions":0.3,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":false},{"name":"Wav2Vec 2.0","article":"https://aiwiki.ai/wiki/wav2vec2","kind":"audio","developer":"Meta (Facebook AI Research)","released":"2020-06-20","parameters":"About 95M (BASE), about 317M (LARGE)","parameters_billions":0.095,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/facebook/wav2vec2-base-960h","article_fact_checked":true},{"name":"Jukebox (OpenAI)","article":"https://aiwiki.ai/wiki/jukebox","kind":"audio","developer":"OpenAI","released":"2020-04-30","parameters":"About 1B (each upsampling prior) to about 5B (top-level prior)","parameters_billions":1,"tier":"noncommercial","weights":"open","license":"custom","license_name":"OpenAI Jukebox Noncommercial Use License","commercial_use":"no","training_code_published":true,"training_data_published":null,"note":"Neither the software nor content created with it may be used commercially.","hugging_face":null,"article_fact_checked":false},{"name":"BlenderBot","article":"https://aiwiki.ai/wiki/blenderbot","kind":"language","developer":"Meta AI (FAIR)","released":"2020-04-29","parameters":"Up to 175B (BlenderBot 3)","parameters_billions":175,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Released through ParlAI (MIT); Hugging Face ports of BlenderBot 1 are tagged Apache 2.0; the BlenderBot 3 175B weights are shared by request.","hugging_face":null,"article_fact_checked":false},{"name":"ColBERT","article":"https://aiwiki.ai/wiki/colbert","kind":"language","developer":"Stanford University (Omar Khattab, Matei Zaharia)","released":"2020-04-27","parameters":"110M (ColBERTv2)","parameters_billions":0.11,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"The official ColBERTv2 checkpoint and the reference code are MIT; third-party ColBERT models carry their own licenses.","hugging_face":"https://huggingface.co/colbert-ir/colbertv2.0","article_fact_checked":true},{"name":"Longformer","article":"https://aiwiki.ai/wiki/longformer","kind":"language","developer":"Allen Institute for AI","released":"2020-04-10","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/allenai/longformer-base-4096","article_fact_checked":true},{"name":"ELECTRA","article":"https://aiwiki.ai/wiki/electra","kind":"language","developer":"Stanford University and Google Brain","released":"2020-03-23","parameters":"14M to 335M","parameters_billions":0.014,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Small, Base and Large checkpoints share the Apache 2.0 license of the google-research/electra repository.","hugging_face":null,"article_fact_checked":true},{"name":"DeBERTa","article":"https://aiwiki.ai/wiki/deberta","kind":"language","developer":"Microsoft Research","released":"2020","parameters":"22M backbone (v3-XSmall) to 1.5B (v2-XXLarge)","parameters_billions":0.022,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"All versions (v1, v2, v3 and multilingual mDeBERTa) are MIT.","hugging_face":null,"article_fact_checked":true},{"name":"DETR","article":"https://aiwiki.ai/wiki/detr","kind":"vision","developer":"Facebook AI Research (FAIR)","released":"2020","parameters":"41M (ResNet-50) to 60M (ResNet-101)","parameters_billions":0.041,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Applies to the original DETR checkpoints; later variants such as Deformable DETR are separate releases.","hugging_face":null,"article_fact_checked":true},{"name":"OPUS-MT","article":"https://aiwiki.ai/wiki/opus-mt","kind":"language","developer":"Language Technology Research Group, University of Helsinki","released":"2020","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"cc-by-4.0","license_name":"CC BY 4.0","commercial_use":"yes","training_code_published":true,"training_data_published":true,"note":"Project models are CC BY 4.0; many Hugging Face ports are tagged Apache 2.0 instead, so check the individual model card.","hugging_face":null,"article_fact_checked":false},{"name":"LayoutLM","article":"https://aiwiki.ai/wiki/layoutlm","kind":"language","developer":"Microsoft Research","released":"2019-12-31","parameters":"~113M (BASE) to ~343M (LARGE), v1","parameters_billions":0.113,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"MIT covers LayoutLM v1 only; LayoutLMv2, LayoutXLM and LayoutLMv3 are CC BY-NC-SA 4.0 (non-commercial).","hugging_face":"https://huggingface.co/microsoft/layoutlm-base-uncased","article_fact_checked":false},{"name":"Sentence-BERT (SBERT)","article":"https://aiwiki.ai/wiki/sentence-bert","kind":"language","developer":"Nils Reimers and Iryna Gurevych, UKP Lab, TU Darmstadt","released":"2019-11","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":true,"note":"Original SBERT checkpoints are Apache 2.0; other models in the sentence-transformers namespace carry their own licenses.","hugging_face":null,"article_fact_checked":true},{"name":"XLM-RoBERTa","article":"https://aiwiki.ai/wiki/xlm-roberta","kind":"language","developer":"Facebook AI Research (Meta AI)","released":"2019-11","parameters":"~270M (Base); ~550M (Large); XL ~3.5B; XXL ~10.7B","parameters_billions":0.27,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/FacebookAI/xlm-roberta-base","article_fact_checked":false},{"name":"T5 (language model)","article":"https://aiwiki.ai/wiki/t5","kind":"language","developer":"Google","released":"2019-10-23","parameters":"60 million to 11 billion (five sizes)","parameters_billions":0.06,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/google-t5/t5-base","article_fact_checked":true},{"name":"DistilGPT2","article":"https://aiwiki.ai/wiki/distilgpt2_model","kind":"other","developer":"Hugging Face","released":"2019-10-03","parameters":"82M (81,912,576 learned parameters)","parameters_billions":0.082,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":true,"note":null,"hugging_face":"https://huggingface.co/distilbert/distilgpt2","article_fact_checked":true},{"name":"BART (language model)","article":"https://aiwiki.ai/wiki/bart","kind":"language","developer":"Facebook AI Research","released":"2019-10","parameters":"~140M (base); ~406M (large)","parameters_billions":0.14,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"The Hugging Face facebook/bart checkpoints are tagged Apache 2.0; the original fairseq release is MIT, which fairseq applies to its pre-trained models.","hugging_face":"https://huggingface.co/facebook/bart-large","article_fact_checked":true},{"name":"DistilBERT","article":"https://aiwiki.ai/wiki/distilbert","kind":"language","developer":"Hugging Face","released":"2019-10","parameters":"66 million","parameters_billions":0.066,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/distilbert/distilbert-base-uncased","article_fact_checked":true},{"name":"ALBERT","article":"https://aiwiki.ai/wiki/albert","kind":"language","developer":"Google Research and Toyota Technological Institute at Chicago","released":"2019-09-26","parameters":"12M to 235M","parameters_billions":0.012,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Four sizes (base to xxlarge, v1 and v2) share the Apache 2.0 license of the google-research/albert repository.","hugging_face":null,"article_fact_checked":true},{"name":"RoBERTa","article":"https://aiwiki.ai/wiki/roberta","kind":"language","developer":"Facebook AI Research; University of Washington","released":"2019-07-26","parameters":"125M (base), 355M (large)","parameters_billions":0.125,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/FacebookAI/roberta-base","article_fact_checked":true},{"name":"XLNet","article":"https://aiwiki.ai/wiki/xlnet","kind":"language","developer":"Carnegie Mellon University and Google Brain","released":"2019-06","parameters":"110M (Base), 340M (Large)","parameters_billions":0.11,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"The official repository (zihangdai/xlnet), which links the released checkpoints, is Apache 2.0; the Hugging Face ports are tagged MIT.","hugging_face":null,"article_fact_checked":true},{"name":"Baidu ERNIE","article":"https://aiwiki.ai/wiki/baidu_ernie","kind":"language","developer":"Baidu, Inc.","released":"2019-04-19","parameters":"Up to 424B total (47B active) in open-source ERNIE 4.5; claimed 2.4T in ERNIE 5.0","parameters_billions":424,"tier":"partial","weights":"partial","license":"custom","license_name":"ERNIE 5.5 Flash Base Model License Agreement v1.0","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":"Latest open release (ERNIE 5.5 Flash Base). The ERNIE 4.5 family (June 2025) is Apache 2.0; ERNIE Bot, 4.0, X1 and 5.0 are API or app only.","hugging_face":null,"article_fact_checked":false},{"name":"GPT-2","article":"https://aiwiki.ai/wiki/gpt-2","kind":"language","developer":"OpenAI","released":"2019-02-14","parameters":"1.5B (XL); also 124M, 355M, 774M variants","parameters_billions":1.5,"tier":"conditional","weights":"open","license":"modified-mit","license_name":"Modified MIT","commercial_use":"yes","training_code_published":false,"training_data_published":false,"note":"Code and checkpoints are under OpenAI's Modified MIT License; WebText and the training code were not released.","hugging_face":"https://huggingface.co/openai-community/gpt2","article_fact_checked":true},{"name":"BioBERT","article":"https://aiwiki.ai/wiki/biobert","kind":"science","developer":"DMIS Lab, Korea University","released":"2019-01-25","parameters":"110M (Base); 340M (Large)","parameters_billions":0.11,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"The dmis-lab/biobert GitHub LICENSE is Apache 2.0; the Hugging Face repositories carry no license tag.","hugging_face":"https://huggingface.co/dmis-lab/biobert-v1.1","article_fact_checked":false},{"name":"CamemBERT","article":"https://aiwiki.ai/wiki/camembert","kind":"language","developer":"Inria (Almanach), Facebook AI Research, Sorbonne Universite","released":"2019","parameters":"110M (base); 335M (large)","parameters_billions":0.11,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/almanach/camembert-base","article_fact_checked":false},{"name":"EfficientNet","article":"https://aiwiki.ai/wiki/efficientnet","kind":"vision","developer":"Google Brain (Mingxing Tan, Quoc V. Le)","released":"2019","parameters":"5.3M (B0) to 66M (B7)","parameters_billions":0.0053,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Official TensorFlow code and checkpoints are Apache 2.0.","hugging_face":null,"article_fact_checked":true},{"name":"SciBERT","article":"https://aiwiki.ai/wiki/scibert","kind":"language","developer":"Allen Institute for AI (AI2)","released":"2019","parameters":"~110M","parameters_billions":0.11,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"The allenai/scibert GitHub LICENSE is Apache 2.0; the Hugging Face repositories carry no license tag.","hugging_face":"https://huggingface.co/allenai/scibert_scivocab_uncased","article_fact_checked":false},{"name":"Wav2Vec","article":"https://aiwiki.ai/wiki/wav2vec","kind":"audio","developer":"Meta AI (formerly Facebook AI Research)","released":"2019","parameters":"About 34M (wav2vec 1.0) to 2B (XLS-R 2B)","parameters_billions":0.034,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Most checkpoints are Apache 2.0; the MMS weights are CC BY-NC 4.0 (non-commercial).","hugging_face":null,"article_fact_checked":true},{"name":"StyleGAN","article":"https://aiwiki.ai/wiki/stylegan","kind":"image","developer":"NVIDIA Research","released":"2018-12-12","parameters":null,"parameters_billions":null,"tier":"noncommercial","weights":"open","license":"cc-by-nc-4.0","license_name":"CC BY-NC 4.0","commercial_use":"no","training_code_published":true,"training_data_published":true,"note":"StyleGAN1 is CC BY-NC 4.0; StyleGAN2, StyleGAN2-ADA and StyleGAN3 use the NVIDIA Source Code License, also non-commercial.","hugging_face":null,"article_fact_checked":true},{"name":"BERT","article":"https://aiwiki.ai/wiki/bert","kind":"language","developer":"Google AI Language","released":"2018-10-11","parameters":"110M (Base); 340M (Large)","parameters_billions":0.11,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/google-bert/bert-base-uncased","article_fact_checked":true},{"name":"Bert-base-uncased model","article":"https://aiwiki.ai/wiki/bert-base-uncased_model","kind":"language","developer":"Google AI Language","released":"2018-10","parameters":"~110 million","parameters_billions":0.11,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/google-bert/bert-base-uncased","article_fact_checked":false},{"name":"GPT","article":"https://aiwiki.ai/wiki/gpt","kind":"language","developer":"OpenAI","released":"2018-06-11","parameters":"117M (GPT-1) to 175B (GPT-3); undisclosed from GPT-4 onward","parameters_billions":0.117,"tier":"partial","weights":"partial","license":"modified-mit","license_name":"Modified MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"GPT-1 (MIT) and GPT-2 (Modified MIT) weights are public; GPT-3 and later are hosted only. gpt-oss (Apache 2.0) has its own article.","hugging_face":null,"article_fact_checked":true},{"name":"GPT-1","article":"https://aiwiki.ai/wiki/gpt-1","kind":"language","developer":"OpenAI","released":"2018-06-11","parameters":"117 million","parameters_billions":0.117,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/openai-community/openai-gpt","article_fact_checked":true},{"name":"ELMo (Embeddings from Language Models)","article":"https://aiwiki.ai/wiki/elmo","kind":"language","developer":"Allen Institute for AI and University of Washington","released":"2018-02","parameters":"~93.6 million","parameters_billions":0.0936,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Checkpoints are hosted by AllenNLP; the official bilm-tf and AllenNLP repos are Apache 2.0, and no separate license is stated for the weights.","hugging_face":null,"article_fact_checked":true},{"name":"CycleGAN","article":"https://aiwiki.ai/wiki/cyclegan","kind":"image","developer":"Berkeley AI Research (BAIR) Lab, UC Berkeley","released":"2017-03","parameters":null,"parameters_billions":null,"tier":"permissive","weights":"open","license":"bsd-3-clause","license_name":"BSD 3-Clause","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Official repo is BSD 2-Clause and distributes the pretrained models; no separate license is stated for the weights.","hugging_face":null,"article_fact_checked":false},{"name":"ResNet","article":"https://aiwiki.ai/wiki/resnet","kind":"vision","developer":"Kaiming He, Xiangyu Zhang, Shaoqing Ren and Jian Sun","released":"2016","parameters":"25.6M (ResNet-50, TorchVision count)","parameters_billions":0.0256,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Architecture page; MIT is the license of the authors' model repository (ResNet-50/101/152). Ports in other libraries carry their own licenses.","hugging_face":null,"article_fact_checked":true},{"name":"YOLO (object detection)","article":"https://aiwiki.ai/wiki/yolo","kind":"vision","developer":"Joseph Redmon and co-authors (v1 to v3); later versions by Ultralytics and other groups","released":"2015-06","parameters":"Varies by version; 55.7M for YOLO26x","parameters_billions":0.055700000000000006,"tier":"conditional","weights":"open","license":"custom","license_name":"GNU AGPL-3.0","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"Copyleft. License of YOLO26, the latest release (Ultralytics also sells an Enterprise license); other versions are mostly GPL-3.0 or AGPL-3.0.","hugging_face":null,"article_fact_checked":true},{"name":"GoogLeNet (Inception v1)","article":"https://aiwiki.ai/wiki/googlenet","kind":"vision","developer":"Google Research","released":"2014-09-17","parameters":"~5M (~6.8M counting all layers incl. auxiliary heads)","parameters_billions":0.005,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":"Google's official TF-Slim Inception v1 checkpoint is distributed on Kaggle Models under Apache 2.0.","hugging_face":null,"article_fact_checked":true},{"name":"VGGNet","article":"https://aiwiki.ai/wiki/vggnet","kind":"vision","developer":"Visual Geometry Group, University of Oxford","released":"2014-09-04","parameters":"138.4M (VGG-16); 143.7M (VGG-19)","parameters_billions":0.1384,"tier":"permissive","weights":"open","license":"cc-by-4.0","license_name":"CC BY 4.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":true},{"name":"d3-turbo","article":"https://aiwiki.ai/wiki/d3_turbo","kind":"vision","developer":"vLLM Semantic Router team","released":null,"parameters":"26.1B loaded (25.6B text backbone, 461M vision encoder, 1.3M answer readout); listed as 27B by the hosted service","parameters_billions":26.1,"tier":"permissive","weights":"open","license":"apache-2.0","license_name":"Apache 2.0","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/vllm-sr/d3-turbo","article_fact_checked":false},{"name":"Ling-3.1-flash","article":"https://aiwiki.ai/wiki/ling_3_1_flash","kind":"language","developer":"InclusionAI (Ant Group)","released":null,"parameters":"About 560B total, 25B active (publisher's count)","parameters_billions":560,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/inclusionAI/Ling-3.1-flash","article_fact_checked":false},{"name":"Ming-3.0","article":"https://aiwiki.ai/wiki/ming_3_0","kind":"audio","developer":"inclusionAI","released":null,"parameters":"7.9B total, about 1.3B active (tts-tiny); 9.3B total, 2.7B active (audio-tiny); 126B total, 6.5B active (audio-flash), per the model cards","parameters_billions":7.9,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":null,"article_fact_checked":false},{"name":"Nemotron-Labs-TwoTower","article":"https://aiwiki.ai/wiki/nemotron_labs_twotower","kind":"language","developer":"NVIDIA","released":null,"parameters":"About 60B total (two 30B towers); about 3B active per tower per token (model card)","parameters_billions":60,"tier":"conditional","weights":"open","license":"nvidia-open-model","license_name":"NVIDIA Open Model License","commercial_use":"conditional","training_code_published":null,"training_data_published":null,"note":null,"hugging_face":"https://huggingface.co/nvidia/Nemotron-Labs-TwoTower-30B-A3B-Base-BF16","article_fact_checked":true},{"name":"Octo (Robot Policy)","article":"https://aiwiki.ai/wiki/octo","kind":"robotics","developer":"Octo Model Team","released":null,"parameters":"27M (Small) and 93M (Base)","parameters_billions":0.027,"tier":"permissive","weights":"open","license":"mit","license_name":"MIT","commercial_use":"yes","training_code_published":true,"training_data_published":null,"note":"MIT covers the code and checkpoints, not every contributed training dataset.","hugging_face":"https://huggingface.co/rail-berkeley/octo-base-1.5","article_fact_checked":true}],"recent_repositories":[{"repository":"https://huggingface.co/nvidia/Kimi-K3.1-Pro-NVFP4","published_by":"nvidia","created_at":"2026-11-11T00:57:46.000Z","model":"Kimi K3.1 Pro","article":"https://aiwiki.ai/wiki/kimi_k3_1_pro"},{"repository":"https://huggingface.co/MiniMaxAI/MiniMax-H3-Image-PE","published_by":"MiniMaxAI","created_at":"2026-11-08T15:35:23.000Z","model":"MiniMax H3","article":"https://aiwiki.ai/wiki/minimax_h3"},{"repository":"https://huggingface.co/tencent/Qwen3.8-27B-DSpine-b16","published_by":"tencent","created_at":"2026-11-07T14:20:29.000Z","model":"Qwen3.8","article":"https://aiwiki.ai/wiki/qwen3_8"},{"repository":"https://huggingface.co/moonshotai/Kimi-K3.1-Pro","published_by":"moonshotai","created_at":"2026-11-02T04:13:51.000Z","model":"Kimi K3.1 Pro","article":"https://aiwiki.ai/wiki/kimi_k3_1_pro"},{"repository":"https://huggingface.co/nvidia/DeepSeek-V4.1-Flash-1017-NVFP4","published_by":"nvidia","created_at":"2026-10-30T17:37:02.000Z","model":"DeepSeek V4.1-Flash","article":"https://aiwiki.ai/wiki/deepseek_v4_1_flash"},{"repository":"https://huggingface.co/tencent/Hy-Memory-30B-A3B-GGUF","published_by":"tencent","created_at":"2026-10-19T07:52:21.000Z","model":"Hy Memory","article":"https://aiwiki.ai/wiki/hy_memory"},{"repository":"https://huggingface.co/MiniMaxAI/MiniMax-M3.1-Flash","published_by":"MiniMaxAI","created_at":"2026-10-17T06:37:50.000Z","model":"MiniMax M3.1 Flash","article":"https://aiwiki.ai/wiki/minimax_m3_1_flash"},{"repository":"https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017","published_by":"deepseek-ai","created_at":"2026-10-17T02:40:38.000Z","model":"DeepSeek V4.1-Flash","article":"https://aiwiki.ai/wiki/deepseek_v4_1_flash"},{"repository":"https://huggingface.co/tencent/Hy-Memory-30B-A3B-FP8","published_by":"tencent","created_at":"2026-10-16T07:57:25.000Z","model":"Hy Memory","article":"https://aiwiki.ai/wiki/hy_memory"},{"repository":"https://huggingface.co/nvidia/Muse-Spark-Open-816B-A42B-DFlash2","published_by":"nvidia","created_at":"2026-10-16T04:40:26.000Z","model":"Muse Spark Open","article":"https://aiwiki.ai/wiki/muse_spark_open"},{"repository":"https://huggingface.co/tencent/Hy-Memory-30B-A3B","published_by":"tencent","created_at":"2026-10-15T09:52:00.000Z","model":"Hy Memory","article":"https://aiwiki.ai/wiki/hy_memory"},{"repository":"https://huggingface.co/nvidia/Nemotron-3.5-Rerank-8B-BF16","published_by":"nvidia","created_at":"2026-10-15T08:42:42.000Z","model":"Nemotron 3","article":"https://aiwiki.ai/wiki/nemotron_3"},{"repository":"https://huggingface.co/stepfun-ai/Step-5-Preview-NVFP4","published_by":"stepfun-ai","created_at":"2026-10-14T02:59:00.000Z","model":"Step 5 Preview","article":"https://aiwiki.ai/wiki/step_5_preview"},{"repository":"https://huggingface.co/nvidia/Qwen-Image-NVFP4","published_by":"nvidia","created_at":"2026-10-09T23:37:43.000Z","model":"Qwen-Image","article":"https://aiwiki.ai/wiki/qwen_image"},{"repository":"https://huggingface.co/Qwen/Qwen-Image-2.1-Turbo","published_by":"Qwen","created_at":"2026-10-09T04:50:01.000Z","model":"Qwen-Image-2.1-Turbo","article":"https://aiwiki.ai/wiki/qwen_image_2_1_turbo"},{"repository":"https://huggingface.co/nvidia/Qwen3.8-Flash-Next-FP4-Local","published_by":"nvidia","created_at":"2026-10-08T16:20:17.000Z","model":"Qwen3.8-Flash-Next","article":"https://aiwiki.ai/wiki/qwen3_8_flash_next"}]}