MultimodalAPI only
Gemini 2.5 Pro
Updated May 21, 2026 · Benchmarks via Artificial Analysis, specs & pricing via OpenRouter · Methodology · Report an error
Superseded by Gemini 3 Pro
Specifications
- Type
- Multimodal
- Access
- API only
- Released
- March 25, 2025
- License
- Proprietary
- Context window
- 1,048,576 tokens
- Knowledge cutoff
- January 31, 2025
- Output speed
- 85 tok/s
- Latency (TTFT)
- 0.7s
- Input
- Text, Image, Audio, Video
- Output
- Text
- API pricing
- $1.25 in · $10.00 out / 1M tokens
Capabilities
Function callingStructured outputWeb search
Benchmarks
Reasoning
Coding
| SWE-bench Verified | 63.8 |
| Aider Polyglot | 76.5 |
| Aider Polyglot Edit | 72.7 |
| LiveCodeBench | 80.1 |
Multimodal
| MMMU | 79.6 |
General
| MMLU-Pro | 86 |
| Humanity’s Last Exam | 17.8 |
| SimpleQA | 50.8 |
Our take
Google’s reasoning flagship, pairing a built-in thinking process with a 1M-token multimodal context. It topped many leaderboards on release and made long-context reasoning a practical default.