Gemini 2.5 Pro Max Thinking
Parameters
-
Context Length
2,097.152K
Modality
Multimodal
Architecture
Dense
License
Proprietary
Release Date
25 Sept 2025
Knowledge Cutoff
-
Technical Specifications
Attention Structure
Multi-Head Attention
Hidden Dimension Size
-
Number of Layers
-
Attention Heads
-
Key-Value Heads
-
Activation Function
-
Normalization
-
Position Embedding
Absolute Position Embedding
Gemini 2.5 Pro Max Thinking
Gemini 2.5 Pro with maximum thinking configuration for complex reasoning tasks. Strong coding (75.69 LiveBench Coding), reasoning (70.81), and mathematics (68.32) performance. Features transparent reasoning process with extended thought chains. Excellent for analytical tasks requiring deep understanding and multi-step problem solving.
About Gemini 2.5
Google's advanced multimodal models with native understanding of text, images, audio, and video. Features massive context windows up to 2.1M tokens, max thinking modes for complex reasoning, and optimized variants for different performance/cost tradeoffs. Includes Pro, Flash, and Flash Lite variants with configurable thinking capabilities for transparent reasoning.
Other Gemini 2.5 Models
Evaluation Benchmarks
Rank
#18
| Benchmark | Score | Rank |
|---|---|---|
QA Assistant ProLLM QA Assistant | 0.97 | 🥈 2 |
Coding Aider Coding | 0.83 | ⭐ 4 |
StackUnseen ProLLM Stack Unseen | 0.83 | 10 |
Coding LiveBench Coding | 0.76 | 15 |
Data Analysis LiveBench Data Analysis | 0.71 | 17 |
Reasoning LiveBench Reasoning | 0.71 | 18 |
Agentic Coding LiveBench Agentic | 0.33 | 26 |
Mathematics LiveBench Mathematics | 0.68 | 32 |
Graduate-Level QA GPQA | 0.86 | 35 |
Rankings
Overall Rank
#18
Coding Rank
#16