Google logo

Gemini 3 Pro

GoogleProprietaryPending Human Review

Gemini 3 Pro is Google's flagship multimodal foundation model, released in November 2025. Built on a sparse Mixture-of-Experts (MoE) Transformer architecture, it features a 1 million token context window and native understanding of text, images, audio, and video. The model introduces 'Deep Think' capabilities for enhanced reasoning, controlled via a 'thinking_level' parameter, and is optimized for 'agentic' workflows and 'vibe coding'—the generation of full applications from natural language. It supports advanced function calling and tool use, making it suitable for complex software engineering and long-context analysis tasks.

2025-11-18
Proprietary
Mixture of Experts (MoE) Transformer
Proprietary

Specifications

Parameters
Proprietary
Architecture
Mixture of Experts (MoE) Transformer
License
Proprietary
Context Window
1,000,000 tokens
Max Output
64,000 tokens
Training Data Cutoff
Jan 2025
Type
multimodal
Modalities
textimageaudiovideo

Benchmark Scores

MMLU-Pro is an enhanced benchmark with over 12,000 challenging questions across 14 domains including...

GPQA91.9

Graduate-level Problems in Quantitative Analysis (GPQA) evaluates advanced reasoning on graduate-lev...

A challenging benchmark of novel problems designed to test the limits of AI capabilities....

American Invitational Mathematics Examination (AIME) 2025 problems....

HMMT95.4

Harvard-MIT Mathematics Tournament (HMMT) problems test advanced high school mathematical problem-so...

A large-scale benchmark of 400 International Mathematical Olympiad-level problems with verifiable an...

Benchmark for evaluating LLMs on code generation tasks from contests....

Software Engineering Benchmark (SWE-bench) evaluates models on real-world software engineering tasks...

Evaluates models on their ability to use terminal commands to solve system administration tasks....

A benchmark for measuring browsing agents' ability to navigate the web and find hard-to-find, entang...

Tool Augmented Understanding Benchmark (TAU-bench) evaluates models on their ability to use tools....

Advanced Specifications

Model Family
Gemini
Finetuned From
null
API Access
Available
Chat Interface
Available
Multilingual Support
Yes
Variants
gemini-3-pro-previewgemini-3-pro-image-preview
Hardware Support
TPU

Capabilities & Limitations

Capabilities
reasoningagenticvibe codingmultimodaltool usefunction callingcodeDeep Think
Known Limitations
hallucinationsmulti-turn degradationlatencyjailbreak vulnerability
Notable Use Cases
complex coding assistantlong-document analysismultimodal agentsvisual reasoning
Function Calling Support
Yes
Tool Use Support
Yes

Related Models