Gemini 2.5 Pro
FrontierMultimodalAPI Available
Benchmark Scores
91.5/100MMLU92
HumanEval93.2
Math86.4
Reasoning92.8
Coding93
About Gemini 2.5 Pro
Gemini 2.5 Pro is Google's most advanced AI model with native thinking capabilities. It leads on coding benchmarks like SWE-bench and offers a 1M token context window with plans for 2M, excelling at complex reasoning and code generation.
Strengths
- +Built-in thinking/reasoning
- +Massive context window
- +Top-tier coding
- +Strong multimodal understanding
Weaknesses
- -Higher latency with thinking enabled
- -Variable output quality
- -Complex pricing tiers
Pricing
Per 1M tokens
Input$1.25
Output$10.00
Quick Facts
- Context Window
- 1M
- Parameters
- Undisclosed
- Release Date
- 2025-03-25
- Category
- Frontier
Compare Models
See how Gemini 2.5 Pro stacks up against other models