Google DeepMind

Gemini 3 Pro

Multimodal
#1τ-bench
#1Humanity's Last Exam
#2GPQA Diamond
+1 more

by Google DeepMind

+
+
+
+
About

Gemini 3 Pro, released by Google in November 2025, represents Google DeepMind's next generation of natively multimodal reasoning — built to extend the Gemini architecture's capabilities in video understanding, complex scientific reasoning, and long-context processing across text, image, and video modalities. It is designed for demanding tasks that require the full breadth of multimodal capability from Google's model family.

+
+
+
+
Pricing Range
Input (per 1M)$2.50 -$2.50
Output (per 1M)$20.00 -$20.00
Providers2
+
+
+
+
Timeline
ReleasedNov 18, 2025
+
+
+
+
Specifications
Capabilities
Multimodal
+
+
+
+
License & Family
License
Proprietary
Performance Overview
Performance metrics and category breakdown

Overall Performance

11 benchmarks
Average Score
57.8%
Best Score
91.9%
High Performers (80%+)
2

Performance Metrics

Max Context Window
1.0M

Top Categories

Science
91.9%
Agents
61.6%
Coding
60.3%
Finance
55.2%
Tool Use
54.1%
+
+
+
+
All Benchmark Results for Gemini 3 Pro
Complete list of benchmark scores with detailed information
GPQA Diamond
Science
91.90
91.9%
Unverified
τ-bench
Agents
85.40
85.4%
Unverified
SWE Bench Verified
Coding
78.00
78.0%
Unverified
BrowseComp
Agents
59.20
59.2%
Unverified
Terminal Bench 2.0
Coding
56.20
56.2%
Unverified
Finance Agent
Finance
55.20
55.2%
Unverified
MCP-Atlas
Tool Use
54.10
54.1%
Unverified
SWE-rebench
Coding
46.70
46.7%
Unverified
GDPVal
Agents
40.30
40.3%
Unverified
Humanity's Last Exam
Reasoning
37.50
37.5%
Unverified
Showing 1 to 10 of 11 benchmarks