ModelBeat
Models/Gemini 2.5 Pro (Jun 2025)
All models
G

Google DeepMind: Gemini 2.5 Pro (Jun 2025)

gemini-2-5-pro-jun-2025

Gemini 2.5 Pro (Jun 2025) is Google DeepMind’s proprietary AI model, released Jun 17, 2025. It scores 145 on the Epoch Capabilities Index.

Google’s general-purpose frontier reasoning model.

Context
—
Input
— / 1M
Output
— / 1M
Released
Jun 17, 2025
Modalities

Benchmarks

Scores on standardized evaluations. Higher is better — and rank shows where Gemini 2.5 Pro (Jun 2025) lands among all models tracked on Model Beat.

15.0
Intelligence Index
Epoch AI
15th percentile of tracked models
3.0
Coding Index
Epoch AI
3rd percentile of tracked models
4.0
Agentic Index
Epoch AI
4th percentile of tracked models

Reasoning

7 evals
ARC-AGI41.0%

Abstract visual reasoning

ARC-AGI-24.9%

Harder abstract reasoning

GPQA Diamond85.3%

Graduate-level scientific reasoning

Humanity's Last Exam21.6%

Frontier of human expert knowledge

SimpleBench62.4%

Common-sense trick questions

SimpleQA Verified56.0%

Factual accuracy & hallucination

WeirdML54.0%

Novel ML problem-solving

Coding

3 evals
GSO (code optimization)3.9%

Code performance optimization

SWE-bench Verified57.6%

Real GitHub issue resolution

WebDev Arena1226

Human-rated web development

Math

3 evals
AIME 2024/202584.2%

Olympiad-qualifier math

FrontierMath14.1%

Research-level math problems

FrontierMath Tier 44.2%

Hardest research math

Agentic & Tools

4 evals
APEX6.6%

Multi-step agentic tasks

GDPval (win/tie rate)25.5%

Economically valuable work

METR task horizon39 min

Autonomous task length

Terminal-Bench32.6%

Command-line agentic tasks

Compare Gemini 2.5 Pro (Jun 2025) with

Frequently asked questions

What is Gemini 2.5 Pro (Jun 2025)?

Google’s general-purpose frontier reasoning model.

Who created Gemini 2.5 Pro (Jun 2025)?

Gemini 2.5 Pro (Jun 2025) is an AI model developed by Google DeepMind, part of the Gemini family, released Jun 17, 2025.

How does Gemini 2.5 Pro (Jun 2025) perform on benchmarks?

On standardized evaluations tracked by Epoch AI, Gemini 2.5 Pro (Jun 2025) scores 56.0% on SimpleQA Verified, 62.4% on SimpleBench, 54.0% on WeirdML.

Is Gemini 2.5 Pro (Jun 2025) open source?

Gemini 2.5 Pro (Jun 2025) is a proprietary model (API access), available through its provider's API rather than as a downloadable open-weight model.

Benchmark data from Epoch AI (CC BY) and Artificial Analysis; pricing, specs & descriptions from OpenRouter. Data updated Jun 16, 2026.