Enjoying AI Stats?Support us
US
Google

Gemini 2.5 Pro Preview

Google
Overview

The latest generally available version of Google's Gemini 2.5 Pro Model. Gemini 2.5 Pro is Google's Flagship multimodal model, offering the strongest performance across any of Google's models with complete multimodaliality, and a 1m input context length.

Quick Links
Key Metrics
Max Input
1,048,576
Tokens
Max Output
65,536
Tokens
Throughput
-
tok/s
Latency
-
ms
Input Price
$1.25
Per 1M Tokens
Cached Input Price
$0.31
Per 1M Tokens
Output Price
$10
Per 1M Tokens
Blended Price
$7.813
Per 1M Tokens
Model Information
Release Details

Released

05 Jun 2025

Knowledge Cutoff

Jan 2025

License

Proprietary

Model Architecture

Parameters

-

Training Data

-

Context Window

Input Context Length

1,048,576 tokens

Output Context Length

65,536 tokens

Key Features
Web Access
Yes

Real-time access to current web information

Multimodal
Yes

Ability to process multiple data types (text, images, etc.)

Reasoning
Yes

Advanced logical and deductive reasoning capabilities

Fine-Tunable
No

Can be customized for specific use cases

Model Release & Updates
17 Jun 2026
Model Deprecated
Model no longer supported or maintained
05 Jun 2025
Model Released
Model first made available to the public
05 Jun 2025
Model Announced
Model first introduced to the public
Benchmarks & Performance Comparison
AI Stats Score
2580.88
#6/191
Ai2 SciArena
1063.00
#3/19
Aider-Polyglot
82.20%
#2/26
AIME 2025
88.00%
#7/27
ARC-AGI-1
41.00%
#5/24
ARC-AGI-2
4.90%
#3/23
Confabulations
12.38
Lower is better
#6/37
Elimation Game
4.04
#25/26
EQ-Bench 3
1476.60
#2/14
GPQA Diamond
86.40%
#3/94
LMArena Text
1470.00
#1/76
LMArena WebDev
1433.16
#1/20
MMLU
89.20%
#2/11
MMMU
82.00%
#1/13
NYT Connections
58.70%
#7/43
SimpleBench
62.40%
#1/31
SimpleQA
54.00%
#1/11
SWE-Bench
67.20%
#2/12
Thematic Generalisation
1.79
Lower is better
#7/34