Compare models
Pick up to 4 models for a side-by-side spec sheet. Hover any dotted metric name for what the test measures. Best value in each row is highlighted.
Gemini 2.5 Pro (batch)GPT-5.6 Sol
| Gemini 2.5 Pro (batch) | GPT-5.6 Sol OpenAI | |
|---|---|---|
| Pricing | ||
| Input $/1M | $0.625 | $2.00 |
| Output $/1M | $5.00 | $10.00 |
| Blended $/1M (3:1) | $1.72 | $4.00 |
| Cache read $/1M | $0.125 | $0.20 |
| Cache write $/1M | — | $2.50 |
| Quality indices | ||
| Intelligence Index | 16.1 | 47 |
| Coding Index | 33.3 | 77.4 |
| Math Index | 87.7 | — |
| Agentic Index | 1.6 | 50.2 |
| Speed & latency | ||
| Output speed (tok/s) | 0 | 65 |
| Time to first token | 0s | 64.46s |
| Time to first answer token | 0s | 64.46s |
| Benchmarks (Artificial Analysis) | ||
| MMLU-Pro | 86.2% | — |
| GPQA Diamond | 84.4% | 94.1% |
| Humanity's Last Exam | 22.5% | 49.5% |
| LiveCodeBench | 80.1% | — |
| SciCode | 46.3% | 57.1% |
| MATH-500 | 96.7% | — |
| AIME | 88.7% | — |
| AIME 2025 | 87.7% | — |
| IFBench | 48.7% | 72.7% |
| AA-LCR (Long Context Reasoning) | 69.0% | 84.0% |
| Terminal-Bench Hard | 26.5% | 65.9% |
| Terminal-Bench 2.1 | 28.5% | 88.0% |
| τ²-Bench (Telecom) | 54.1% | 85.1% |
| τ³-Bench Banking | 9.7% | 44.3% |
| Design Arena (Elo) | ||
| Website | 1,179 | — |
| UI components | 1,155 | — |
| Data viz | 1,236 | — |
| 3D | 1,109 | — |
| Game dev | 1,135 | — |
| Code categories | 1,167 | — |
| Specs | ||
| Context window | 1.05M | 1.05M |
| Max output tokens | 66K | 128K |
| Input modalities | text, image, file, audio, video | file, image, text |
| Tokenizer | Gemini | GPT |