← All comparisons
Gemini 3 Flash Preview (Non-reasoning) vs Claude 4.1 Opus (Reasoning)
Google vs Anthropic — side-by-side benchmark comparison
| Gemini 3 Flash Preview (Non-reasoning) | Claude 4.1 Opus (Reasoning) | |
|---|---|---|
| Intelligence Index | 35.0 | 42.0 |
| Coding Index | 37.8 | 36.5 |
| Math Index | 55.7 | 80.3 |
| Output speed (tok/s) | 189.8 | 44.5 |
| Blended price ($/1M) | $1.13 | $32.81 |
| Time to first token (s) | 0.90s | 8.55s |
| aime | — | — |
| aime 25 | 55.7% | 80.3% |
| artificial analysis coding index | 37.80 | 36.50 |
| artificial analysis intelligence index | 35.00 | 42.00 |
| artificial analysis math index | 55.70 | 80.30 |
| gpqa | 81.2% | 80.9% |
| hle | 14.1% | 11.9% |
| ifbench | 55.1% | 55.4% |
| lcr | 48.0% | 66.3% |
| livecodebench | 79.7% | 65.4% |
| math 500 | — | — |
| mmlu pro | 88.2% | 88.0% |
| scicode | 49.9% | 40.9% |
| tau2 | 43.3% | 71.4% |
| terminalbench hard | 31.8% | 34.3% |
Benchmark data from Artificial Analysis.