Side-by-side comparison of AI model specs, benchmarks, and pricing.
| Spec | Claude Sonnet 4.5 Anthropic | Gemini 2.5 Flash Google |
|---|---|---|
| Tier | frontier | mid |
| Release Date | 2025-09-29 | 2025-06-17 |
| Context Window | 200K | 1049K |
| Max Output | 64K | 66K |
| Input Price (/1M) | $3 | $0.3 |
| Output Price (/1M) | $15 | $2.5 |
| Arena Elo | 1,450 | 1,410 |
| MMLU | 88.7% | 83.5% |
| GPQA | 83.4% | 65.8% |
| MATH | 87% | 82.1% |
| HumanEval | 95% | 90.3% |
| SWE-bench | 55.8% | — |
| AIME | 58% | 58% |
| SimpleQA | 30.8% | 28.3% |
| Capabilities | vision, tool-use, code, extended-thinking | vision, tool-use, code, reasoning, audio |