Thursday, July 23, 2026
- Qwen3-Coderv3Qwen3-Coder is a family of open-weight large language models specialized for software engineering, developed by Alibaba's Qwen team (Tongyi Lab) and released under the Apache 2.0 license.
- GPT-5 Codexv3GPT-5-Codex is a coding-specialised variant of OpenAI's GPT-5 model, announced on September 15, 2025 and tuned for agentic software engineering inside the OpenAI Codex product family.
- Thinking Machines Labv7Thinking Machines Lab is an American artificial intelligence research company, headquartered in San Francisco, California, that was founded in February 2025 by Mira Murati, the former chief technology officer…
- Decartv4Decart (also known as Decart AI, operating from the domain decart.ai) is an Israeli artificial intelligence company headquartered in Tel Aviv that builds real-time generative AI systems for video and…
- Voyage-3v3Voyage-3 is a family of general-purpose text embedding models developed by Voyage AI, launched in September 2024 with voyage-3 and voyage-3-lite , expanded in January 2025 with voyage-3-large , and refreshed…
- Jina Embeddings v3v3Jina Embeddings v3 is a multilingual text embedding model released by Jina AI on September 18, 2024, with 570 million parameters, support for 89 languages, an 8,192 token context window, and a stack of…
- Agentic RAGv3Agentic RAG (agentic retrieval-augmented generation) is a retrieval-augmented generation design pattern in which one or more autonomous large language model AI agents plan, execute, and revise a sequence of…
- Process reward model (PRM)v6A process reward model (PRM), also called a process-supervised reward model or step-level verifier, is a learned scoring model that evaluates the correctness or quality of each intermediate step in a large…
- RLAIFv6Reinforcement Learning from AI Feedback (RLAIF) is a family of alignment techniques for large language models in which the preference labels used to fine-tune a model are produced by another AI system
- Proximal Policy Optimization (PPO)v4Proximal Policy Optimization (PPO) is an on-policy policy gradient reinforcement learning algorithm that stabilizes training by clipping the policy update so the new policy stays close ("proximal") to the old…
- Matryoshka representation learningv3Matryoshka Representation Learning (MRL) is a representation learning technique that trains a single neural model to produce embedding vectors which remain useful when truncated to many smaller dimensionalities
- Deepgramv4Deepgram is an American voice artificial intelligence company, founded in 2015 and headquartered in San Francisco, that builds proprietary deep learning models for speech recognition, text-to-speech synthesis…
- Common Corpusv3Common Corpus is the largest fully open, multilingual dataset for pretraining large language models, assembled and released by the French AI research lab Pleias.
- Speechmaticsv3Speechmatics is a British artificial intelligence company that develops automatic speech recognition (ASR) and voice AI technology for enterprise customers.
- Blackhole (Tenstorrent)v3Blackhole is the third-generation AI accelerator architecture from Tenstorrent, the Toronto and Santa Clara based fabless semiconductor company led by CEO Jim Keller.
- DCLM (DataComp for Language Models)v3DCLM, short for DataComp for Language Models (also styled DataComp-LM), is an open benchmark, dataset, and software framework, released in June 2024
- Common Pilev4Common Pile v0.1 is an 8 terabyte corpus of openly licensed and public domain text, released on June 5, 2025, by EleutherAI and a consortium of more than two dozen academic and industry collaborators.
- Abilene data center (Stargate)v3The Abilene data center, also called the Crusoe Abilene Stargate Campus or Stargate I, is the flagship artificial intelligence data center of the Stargate Project, built by Crusoe Energy on a roughly 1,000…
- Abacus.AIv2Abacus.AI is an American enterprise artificial intelligence and machine learning platform company headquartered in San Francisco, California.
- LMDeployv4LMDeploy is an open-source toolkit for compressing, deploying, and serving large language models, developed by the MMRazor and MMDeploy teams associated with the InternLM project at the Shanghai AI Laboratory.
- Pleiasv3Pleias (stylized PleIAs) is a Paris based artificial intelligence laboratory and small company that designs, pretrains, and releases large language models trained exclusively on public domain and permissively…
- GPAI Code of Practicev3The General-Purpose AI Code of Practice (abbreviated GPAI Code of Practice or CoP) is a voluntary compliance framework published by the European Commission through its AI Office on 10 July 2025 to help…
- Contextual AIv5Contextual AI is an American enterprise artificial intelligence company headquartered in Mountain View, California, that builds production-grade systems based on retrieval-augmented generation.
- Bret Taylorv3Bret Steven Taylor (born 1980) is an American software engineer and technology executive who serves as chairman of the board of OpenAI and as co-founder and chief executive officer of Sierra
- Nemotron-CCv4Nemotron-CC is a large-scale, open English-language pretraining dataset for large language models released by NVIDIA in December 2024.
- Spatial intelligencev3Spatial intelligence is the ability of an AI system to perceive, understand, reason about, generate, and interact with three-dimensional space rather than just text or two-dimensional pixels.
- Tenstorrentv4Tenstorrent is a North American artificial intelligence hardware and intellectual property company that designs processors for AI training and inference on the basis of the open-standard RISC-V instruction set…
- OpenPIv4OpenPI (stylized openpi) is the open-source repository of robot foundation models, training code, and inference utilities published by Physical Intelligence, the San Francisco robotics and AI startup…
- Marble (World Labs)v4Marble is a multimodal generative world model developed by World Labs, the spatial intelligence startup co-founded by Stanford computer scientist Fei-Fei Li.
- IsoDDEv4IsoDDE, short for Isomorphic Labs Drug Design Engine, is a unified computational drug design system developed by Isomorphic Labs, the Alphabet subsidiary spun out of Google DeepMind in 2021.
- V-JEPA 2v4V-JEPA 2 (Video Joint Embedding Predictive Architecture 2) is an open-source video world model released by Meta AI on June 11, 2025 that learns to understand, predict, and plan in the physical world by…
- ESM3v3ESM3 (Evolutionary Scale Modeling 3) is a frontier multimodal generative language model for biology, released by EvolutionaryScale on June 25, 2024, that was the first model to reason jointly over the…
- EvolutionaryScalev5EvolutionaryScale is an American artificial intelligence company that builds frontier generative models for biology, best known for ESM3, a multimodal generative protein language model that can reason over and…
- Gemini 2.5 Flashv5Gemini 2.5 Flash is a fast, cost-optimized multimodal large language model developed by Google DeepMind and the mid-tier member of the Gemini 2.5 family.
- V-JEPAv3V-JEPA (Video Joint Embedding Predictive Architecture) is a self-supervised video model from Meta AI that learns by predicting masked regions of a video in an abstract latent representation space rather than…
- LINGO-2 (Wayve)v3LINGO-2 is a closed-loop vision-language-action model for autonomous driving developed by the British self-driving company Wayve.
- AlphaGenomev3AlphaGenome is a unified deep learning model developed by Google DeepMind that predicts thousands of functional genomic properties from raw DNA sequences.
- SmolLM 3v3SmolLM 3 is a fully open 3 billion parameter language model released by Hugging Face on July 8, 2025, trained on 11.2 trillion tokens and designed as a small, multilingual, long-context reasoner.
- Jet-Nemotronv3Jet-Nemotron is a family of small hybrid-architecture language models released by NVIDIA Research in August 2025.
- RWKV-7 (Goose)v3RWKV-7, codenamed Goose, is an attention-free, RNN-style large-language-model architecture introduced in March 2025 that runs inference in linear time with constant memory per token while still training in…
- Zyphrav4Zyphra is an American artificial intelligence research and product company headquartered in San Francisco, California, with a secondary office in London.
- ZAYA1-8Bv4ZAYA1-8B is an open-weight, reasoning-focused Mixture-of-Experts (MoE) large language model released by San Francisco-based AI research lab Zyphra on May 6, 2026.
- Retentive Network (RetNet)v4RetNet (Retentive Network) is a sequence-modeling architecture proposed by Microsoft Research and Tsinghua University in July 2023 as a successor to the Transformer for large language models.
- Jamba2v4Jamba2 is the second generation of hybrid State Space Model and Transformer language models released by AI21 Labs on January 8, 2026.
- Llama 4 Behemothv3Llama 4 Behemoth is the announced but never publicly released flagship model in the Llama 4 family from Meta AI.
- Mixtral 8x22Bv6Mixtral 8x22B is a sparse mixture-of-experts (MoE) large language model released by the French AI company Mistral AI on April 17, 2024.
- OpenAI Codex Cloudv2OpenAI Codex Cloud is a hosted, cloud-based software engineering agent developed by OpenAI and launched as a research preview on May 16, 2025.
- Reka Corev3Reka Core is a frontier class multimodal foundation model developed by Reka AI, a research and product company founded in 2022 by former scientists from DeepMind, Google Brain, Meta FAIR, and Baidu.
- Reka Flashv3Reka Flash is a family of multimodal large language models developed by Reka AI, a San Francisco Bay Area research company founded in 2022 by former researchers from Google DeepMind, Meta FAIR, and Google.
- Reka Edgev3Reka Edge is a 7-billion-parameter multimodal language model developed by Reka AI, introduced in April 2024 as the smallest member of the company's first publicly described model family.
- Tülu 3v4Tülu 3 is a fully open post-training recipe and a corresponding family of instruction-tuned language models released by the Allen Institute for AI (Ai2) on November 21, 2024.
- OLMo 3v4OLMo 3 is the third generation of fully open language models released by the Allen Institute for AI (Ai2).
- OLMo 2v3OLMo 2 is the second generation of fully open large language models released by the Allen Institute for AI (Ai2), spanning 7B, 13B, and 32B parameter sizes.
- Tripo P1v3Tripo P1, marketed in full as Tripo Smart Mesh P1.0, is a production grade native 3D diffusion model that generates clean, engine ready 3D meshes from text or image prompts in as little as two seconds.
- Rodin Gen-2v3Rodin Gen-2 is a generative artificial intelligence model for producing three dimensional assets from text prompts or reference images, developed by the Shanghai based company Deemos and delivered through the…
- Phindv2Phind was an AI-powered answer engine built for software developers. The product combined a live web index with fine-tuned large language models to return cited, code-aware answers to programming questions…
- SmolLMv4SmolLM is a family of small, fully open language models released by Hugging Face on July 16, 2024 in three sizes, 135 million, 360 million, and 1.7 billion parameters, all trained on a curated open dataset…
- Microsoft Foundry Localv3Foundry Local is an on-device artificial intelligence runtime from Microsoft that lets applications run open weight language models entirely on a user's own hardware.
- OLMoEv3OLMoE (Open Mixture-of-Experts) is a fully open sparse mixture of experts large language model released by the Allen Institute for AI (Ai2) on September 3, 2024 .
- Sora (app)v4Sora is a short-form video application released by OpenAI on September 30, 2025, alongside the Sora 2 text-to-video model that powers it.
- Hedra Characterv5Hedra Character is a family of generative video foundation models that turn a single image plus an audio clip (and, in later versions, a text prompt) into a video in which the pictured person or character…
- SmolLM 2v4SmolLM 2 is a family of compact open-weight language models released by Hugging Face on November 1, 2024.
- HeyGen Avatar IVv4Avatar IV is the fourth generation of the AI avatar engine from HeyGen, the AI video company co-founded in 2020 by Joshua Xu and Wayne Liang.
- Meshy 6v4Meshy 6 is the sixth major release of Meshy AI's generative 3D generation platform, a hosted generative AI service that turns text prompts or images into textured 3D models (meshes) for games, animation, and…
- Claude Code Subagentsv3Claude Code subagents (also sub-agents) are specialized AI assistants that Claude Code, Anthropic's command-line coding tool, can delegate specific tasks to.
- Phi-4-miniv3Phi-4-mini is a 3.8 billion parameter open weight small language model released by Microsoft on February 26, 2025, under the permissive MIT license.
- Runway Alephv3Runway Aleph is an in-context AI video editing model from Runway that transforms and edits an existing video clip from a plain-text instruction, performing a wide range of tasks in a single model: adding…
- Runway Act-Twov3Runway Act-Two is a generative motion capture and character animation model developed by Runway, publicly introduced on July 15, 2025.
- GitHub Sparkv2GitHub Spark is a natural-language application builder developed by GitHub that lets users describe a web app in plain English and watch it be generated, deployed, and hosted without manually writing or…
- Genesis (simulator)v3Genesis is an open-source, generative physics simulation platform for robotics and embodied AI, released on December 19, 2024 after a roughly two-year (24-month) collaboration involving more than 20 academic…
- OpenAI AgentKitv5OpenAI AgentKit is a suite of agent-building tools that OpenAI introduced at OpenAI DevDay on 6 October 2025 to take AI agents from prototype to production on OpenAI's hosted models.
- Phi-4 Reasoningv4Phi-4-reasoning is a 14 billion parameter open weight reasoning model released by Microsoft Research on April 30, 2025.
- Pika 2.5v3Pika 2.5 is a generative video model developed by Pika Labs, the San Francisco based AI video startup co-founded in April 2023 by Stanford AI Lab dropouts Demi Guo and Chenlin Meng.
- Phi-4-mini-flash-reasoningv4Phi-4-mini-flash-reasoning is a 3.8 billion parameter open weight reasoning model released by Microsoft in July 2025.
- BYD humanoid (BoYoboD)v3The BYD humanoid, reported under the name "BoYoboD," is a widely circulated story about a $10,000 solar-powered home robot from Chinese electric vehicle giant BYD
- ALOHA 2v4ALOHA 2 is an open-source, low-cost bimanual teleoperation hardware platform released in February 2024 by an Google DeepMind led team working with the original ALOHA authors at Stanford University.
- Intel RealSense D555v4The Intel RealSense D555 (also branded simply as RealSense D555 PoE) is a stereoscopic depth camera announced in mid-2025 that is the first product in the RealSense D400 family to integrate Power over Ethernet…
- Synthesia 3.0v4Synthesia 3.0 is a major release of the AI video generation platform from Synthesia, the London-based company co-founded in 2017 by Victor Riparbelli, Steffen Tjerrild, Lourdes Agapito and Matthias Niessner.
- Wan 2.5v5Wan 2.5 is a natively multimodal AI video generation model developed by Alibaba Cloud's Tongyi Lab and previewed at the company's Apsara 2025 conference in Hangzhou on September 24, 2025.
- Sesame CSMv3Sesame CSM (Conversational Speech Model) is an open weights speech generation model from Sesame AI, a San Francisco startup co-founded by former Oculus chief executive Brendan Iribe.
- Gemini 3 Flashv4Gemini 3 Flash is a multimodal large language model released by Google on December 17, 2025 as the fast, lower-cost sibling to Gemini 3 Pro in the Gemini 3 family.
- Persona AIv4Persona AI is an American robotics company headquartered in Houston, Texas, that designs industrial humanoid robots for physically demanding work in shipyards, factories, construction sites, and energy…
- Falcon 3v4Falcon 3 is a family of open-weight large language models released on December 17, 2024 by the Technology Innovation Institute (TII), an applied research center based in Abu Dhabi, United Arab Emirates .
- Cohere Command Av6Cohere Command A is a 111 billion parameter dense large language model released by Cohere on March 13, 2025, built for enterprise agents, Retrieval-Augmented Generation, and tool use across 23 languages.
- Reflex Robotics humanoidv3The Reflex Robotics humanoid is a wheeled, dual-arm general purpose robot built by Reflex Robotics, a New York City startup founded in 2022 by a small team of engineers with prior experience at Boston…
- Perplexity Financev2Perplexity Finance is a dedicated financial research product from Perplexity AI, available at perplexity.ai/finance, that combines real-time market data, company filings, earnings call transcripts, and…
- Hume Octave 2v3Hume Octave 2 is a multilingual emotional text-to-speech model released by Hume AI on October 1, 2025.
- ElevenLabs v3v4Eleven v3, marketed by ElevenLabs as Eleven v3 (alpha), is a third-generation text-to-speech model that ElevenLabs released in public alpha on June 5, 2025 and described as "the most expressive Text to Speech…
- Mistral Large 3v7Mistral Large 3 is a sparse mixture-of-experts large language model released on December 2, 2025 by the French AI company Mistral AI, distributed as open weights under the Apache 2.0 license with roughly 675…
- Stable Audio 2.5v4Stable Audio 2.5 is an enterprise focused text-to-audio generation model released by Stability AI on September 10, 2025.
- Agnov3Agno is an open-source Python framework for building, running, and managing AI agents and multi-agent systems.
- Replit Agentv6Replit Agent is an AI software development agent built by Replit that turns a natural language description into a deployed, full-stack web application without the user writing code.
- Hunyuan 3Dv4Hunyuan 3D is a family of open weight generative artificial intelligence models from Tencent that turn text prompts, single images, sketches, and other inputs into ready to use three dimensional assets…
- Wan 2.1-VACEv3Wan 2.1-VACE (also written Wan2.1-VACE) is an open-weights video creation and editing model released by Alibaba's Tongyi Lab on May 14, 2025 .
- Atlas Electric (Boston Dynamics)v5Atlas Electric is the all-electric, production-grade humanoid robot designed and manufactured by Boston Dynamics.
- Yi-Lightningv3Yi-Lightning is a closed-source large language model developed by Chinese artificial intelligence company 01.AI (零一万物, Língyī Wànwù), the company founded by Kai-Fu Lee.
- Kimi K2.5v4Kimi K2.5 is an open-weights, natively multimodal large language model developed by Moonshot AI and released on January 27, 2026 .
- Yi-Largev3Yi-Large is a closed-source large language model developed by Chinese artificial intelligence company 01.AI (零一万物, Língyi Wànwù), founded by Kai-Fu Lee.
- Mastrav3Mastra is an open-source TypeScript framework for building AI agents, workflows, and retrieval-augmented generation pipelines.
- Seedancev6Seedance is the family of foundation video generation models built by the Seed team at ByteDance, the Chinese internet company that owns TikTok and Douyin.
