← All comparisons
Gemini 2.5 Pro vs Claude 4.1 Opus (Reasoning)
Google vs Anthropic — side-by-side benchmark comparison
| Gemini 2.5 Pro | Claude 4.1 Opus (Reasoning) | |
|---|---|---|
| Intelligence Index | 34.6 | 42.0 |
| Coding Index | 32.0 | 36.5 |
| Math Index | 87.7 | 80.3 |
| Output speed (tok/s) | 140.4 | 44.5 |
| Blended price ($/1M) | $3.44 | $32.81 |
| Time to first token (s) | 20.87s | 8.55s |
| aime | 88.7% | — |
| aime 25 | 87.7% | 80.3% |
| artificial analysis coding index | 32.00 | 36.50 |
| artificial analysis intelligence index | 34.60 | 42.00 |
| artificial analysis math index | 87.70 | 80.30 |
| gpqa | 84.4% | 80.9% |
| hle | 21.1% | 11.9% |
| ifbench | 48.7% | 55.4% |
| lcr | 66.0% | 66.3% |
| livecodebench | 80.1% | 65.4% |
| math 500 | 96.7% | — |
| mmlu pro | 86.2% | 88.0% |
| scicode | 42.8% | 40.9% |
| tau2 | 54.1% | 71.4% |
| terminalbench hard | 26.5% | 34.3% |
Benchmark data from Artificial Analysis.