← All comparisons
Gemini 2.5 Flash (Non-reasoning) vs Claude 4.1 Opus (Reasoning)
Google vs Anthropic — side-by-side benchmark comparison
| Gemini 2.5 Flash (Non-reasoning) | Claude 4.1 Opus (Reasoning) | |
|---|---|---|
| Intelligence Index | 20.6 | 42.0 |
| Coding Index | 17.8 | 36.5 |
| Math Index | 60.3 | 80.3 |
| Output speed (tok/s) | 214.5 | 44.5 |
| Blended price ($/1M) | $0.85 | $32.81 |
| Time to first token (s) | 0.51s | 8.55s |
| aime | 50.0% | — |
| aime 25 | 60.3% | 80.3% |
| artificial analysis coding index | 17.80 | 36.50 |
| artificial analysis intelligence index | 20.60 | 42.00 |
| artificial analysis math index | 60.30 | 80.30 |
| gpqa | 68.3% | 80.9% |
| hle | 5.1% | 11.9% |
| ifbench | 39.0% | 55.4% |
| lcr | 45.9% | 66.3% |
| livecodebench | 49.5% | 65.4% |
| math 500 | 93.2% | — |
| mmlu pro | 80.9% | 88.0% |
| scicode | 29.1% | 40.9% |
| tau2 | 14.9% | 71.4% |
| terminalbench hard | 12.1% | 34.3% |
Benchmark data from Artificial Analysis.