CtrlK
Gemini 3 Pro is a preview flagship model in the Gemini family from Google DeepMind. Google introduced Gemini 3 on November 18, 2025 and said it was beginning the Gemini 3 era by releasing Gemini 3 Pro in preview across the Gemini app, Google AI Studio, Vertex AI, and related products.[1][2]
Google's current public product page is titled "Gemini 3.1 Pro," but the model information on that page still identifies the preview line as "3 Pro." Google presents it as the best Gemini model for complex tasks, advanced coding, long-context work, and multimodal understanding.[2]
| Field | Public details from Google |
|---|---|
| Status | Preview |
| Input types | Text, image, video, audio, and PDF |
| Output | Text |
| Context window | 1 million input tokens |
| Output limit | 64k tokens |
| Knowledge cutoff | January 2025 |
| Tool use | Function calling, structured output, search as a tool, code execution |
| Availability | Gemini app, Google AI Studio, Vertex AI, Gemini API, Google AI Mode, Google Antigravity |
Google's November 2025 announcement described Gemini 3 Pro as state of the art on a wide range of reasoning and multimodal benchmarks.[1][3]
| Evaluation | Result cited by Google |
|---|---|
| Humanity's Last Exam | 37.5% without tools |
| GPQA Diamond | 91.9% |
| MathArena Apex | 23.4% |
| MMMU-Pro | 81.0% |
| Video-MMMU | 87.6% |
| SimpleQA Verified | 72.1% |
| Terminal-Bench 2.0 | 54.2% |
| WebDev Arena | 1487 Elo |
| SWE-bench Verified | 76.2% |
Google's developer blog said Gemini 3 Pro launched in preview at $2 per million input tokens and $12 per million output tokens for prompts of 200,000 tokens or less through the Gemini API in Google AI Studio and Vertex AI.[3]
The same launch materials describe Gemini 3 Pro as Google's best model for vibe coding and agentic coding, and say it fits into existing production coding and agent workflows while also powering the company's Google Antigravity development platform.[1][3]
By early 2026 Google DeepMind's main model page had shifted to the label "Gemini 3.1 Pro" while preserving the same core positioning: a top-tier Gemini 3-series model for complex, multimodal, and agentic tasks. That makes Gemini 3 Pro and Gemini 3.1 Pro best understood as successive public presentations of the same flagship preview line rather than unrelated products.[2][4]