Google DeepMind DeepMind
code
Build with Gemini
spark
Try Gemini
Gemini 3 Pro

Best for complex tasks and bringing creative concepts to life

Try in Gemini
Build with Gemini

Our most intelligent model yet. With state-of-the-art reasoning to help you learn, build, and plan anything.

Partner with a pro

With state-of-the-art reasoning and multimodal capabilities

Capabilities

Reasoning with unprecedented depth and nuance

Smart, concise, direct responses – with genuine insight over cliche and flattery.

World-leading multimodal understanding

Text, images, video, audio – even code. Gemini 3 is state-of-the-art on reasoning with unprecedented depth and nuance.

Our best model for vibe coding and agentic coding

Gemini 3 brings exceptional instruction following – with meaningful improved tool use and agentic coding.

Improved agentic capabilities

Better tool use. Simultaneous, multi-step tasks. Gemini 3’s agentic capabilities can build more helpful and intelligent personal AI assistants.

Get started

Build with Gemini 3

New

Google Antigravity

Build with our new agentic development platform

Download
Learn more

Google AI Studio

Leap from prompt to production

Try Google AI Studio

Gemini API

Get started building with cutting-edge AI models

Build with Gemini

Hands-on

Explore what you can do with Gemini 3 Pro

Vibe code like never before to create beautiful, responsive experiences

Gemini 3 Pro excels at practical, front-end development – with a more intuitive interface and richer design

Gain a deeper understanding through more nuanced responses

Gemini 3 Pro’s state-of-the-art reasoning provides unprecedented nuance and depth

Generate code for interactive flashcards, games and experiences to help you master new material

Gemini 3 Pro seamlessly synthesizes information across text, images, video, audio, and even code to help you learn

Performance

Gemini 3 is state-of-the-art across a wide range of benchmarks

Our most intelligent model yet sets a new bar for AI model performance.

Bar chart titled "Humanity’s Last Exam" measuring reasoning and knowledge. Gemini 3 Pro achieves the highest score at 37.5%, followed by GPT-5.1 at 26.5%, Gemini 2.5 at 21.6%, and Claude Sonnet 4.5 at 13.7%. Bar chart titled "Humanity’s Last Exam" measuring reasoning and knowledge. Gemini 3 Pro achieves the highest score at 37.5%, followed by GPT-5.1 at 26.5%, Gemini 2.5 at 21.6%, and Claude Sonnet 4.5 at 13.7%.
Bar chart titled "Terminal-Bench 2.0" measuring agentic coding. Gemini 3 Pro scores highest at 54.2%, followed by GPT-5.1 at 47.6%, Claude Sonnet 4.5 at 42.8%, and Gemini 2.5 at 32.6%. Bar chart titled "Terminal-Bench 2.0" measuring agentic coding. Gemini 3 Pro scores highest at 54.2%, followed by GPT-5.1 at 47.6%, Claude Sonnet 4.5 at 42.8%, and Gemini 2.5 at 32.6%.
Bar chart titled "SimpleQA Verified" measuring parametric knowledge. Gemini 3 Pro shows a significant lead with 72.1%, followed by Gemini 2.5 at 54.5%, GPT-5.1 at 34.9%, and Claude Sonnet 4.5 at 29.3%. Bar chart titled "SimpleQA Verified" measuring parametric knowledge. Gemini 3 Pro shows a significant lead with 72.1%, followed by Gemini 2.5 at 54.5%, GPT-5.1 at 34.9%, and Claude Sonnet 4.5 at 29.3%.
Benchmark Notes Gemini 3 Pro Gemini 2.5 Pro Claude Sonnet 4.5 GPT-5.1
Academic reasoning Humanity's Last Exam No tools 37.5% 21.6% 13.7% 26.5%
With search and code execution 45.8% — — —
Visual reasoning puzzles ARC-AGI-2 ARC Prize Verified 31.1% 4.9% 13.6% 17.6%
Scientific knowledge GPQA Diamond No tools 91.9% 86.4% 83.4% 88.1%
Mathematics AIME 2025 No tools 95.0% 88.0% 87.0% 94.0%
With code execution 100.0% — 100.0% —
Challenging Math Contest problems MathArena Apex 23.4% 0.5% 1.6% 1.0%
Multimodal understanding and reasoning MMMU-Pro 81.0% 68.0% 68.0% 76.0%
Screen understanding ScreenSpot-Pro 72.7% 11.4% 36.2% 3.5%
Information synthesis from complex charts CharXiv Reasoning 81.4% 69.6% 68.5% 69.5%
OCR OmniDocBench 1.5 Overall Edit Distance, lower is better 0.115 0.145 0.145 0.147
Knowledge acquisition from videos Video-MMMU 87.6% 83.6% 77.8% 80.4%
Competitive coding problems LiveCodeBench Pro Elo Rating, higher is better 2,439 1,775 1,418 2,243
Agentic terminal coding Terminal-Bench 2.0 Terminus-2 agent 54.2% 32.6% 42.8% 47.6%
Agentic coding SWE-Bench Verified Single attempt 76.2% 59.6% 77.2% 76.3%
Agentic tool use τ2-bench 85.4% 54.9% 84.7% 80.2%
Long-horizon agentic tasks Vending-Bench 2 Net worth (mean), higher is better $5,478.16 $573.64 $3,838.74 $1,473.43
Held out internal grounding, parametric, MM, and search retrieval benchmarks FACTS Benchmark Suite 70.5% 63.4% 50.4% 50.8%
Parametric knowledge SimpleQA Verified 72.1% 54.5% 29.3% 34.9%
Multilingual Q&A MMMLU 91.8% 89.5% 89.1% 91.0%
Commonsense reasoning across 100 Languages and Cultures Global PIQA 93.4% 91.5% 90.1% 90.9%
Long context performance MRCR v2 (8-needle) 128k (average) 77.0% 58.0% 47.1% 61.6%
1M (pointwise) 26.3% 16.4% not supported not supported

For details on our evaluation methodology please see deepmind.google/models/evals-methodology/gemini-3-pro

Model information

Name
3 Pro
Status
Preview
Input
  • Text
  • Image
  • Video
  • Audio
  • PDF
Output
  • Text
Input tokens
1M
Output tokens
64k
Knowledge cutoff
January 2025
Tool use
  • Function calling
  • Structured output
  • Search as a tool
  • Code execution
Best for
  • Agentic
  • Advanced coding
  • Long context understanding
  • Multimodal understanding
  • Algorithmic development
Availability
  • Gemini App
  • Google Cloud / Vertex AI
  • Google AI Studio
  • Gemini API
  • Google AI Mode
  • Google Antigravity
Documentation
View developer docs
Model card
View model card