← All comparisons
Gemini 2.5 Flash Preview (Sep '25) (Non-reasoning) vs Claude 4.5 Sonnet (Reasoning)
Google vs Anthropic — side-by-side benchmark comparison
| Gemini 2.5 Flash Preview (Sep '25) (Non-reasoning) | Claude 4.5 Sonnet (Reasoning) | |
|---|---|---|
| Intelligence Index | 25.7 | 43.0 |
| Coding Index | 22.1 | 38.6 |
| Math Index | 56.7 | 88.0 |
| Output speed (tok/s) | 0.0 | 55.0 |
| Blended price ($/1M) | $0.00 | $6.56 |
| Time to first token (s) | 0.00s | 7.02s |
| aime | — | — |
| aime 25 | 56.7% | 88.0% |
| artificial analysis coding index | 22.10 | 38.60 |
| artificial analysis intelligence index | 25.70 | 43.00 |
| artificial analysis math index | 56.70 | 88.00 |
| gpqa | 76.6% | 83.4% |
| hle | 7.8% | 17.3% |
| ifbench | 43.5% | 57.3% |
| lcr | 56.7% | 65.7% |
| livecodebench | 62.5% | 71.4% |
| math 500 | — | — |
| mmlu pro | 83.6% | 87.5% |
| scicode | 37.5% | 44.7% |
| tau2 | 28.4% | 78.1% |
| terminalbench hard | 14.4% | 35.6% |
Benchmark data from Artificial Analysis.