← All comparisons
Claude 4.5 Haiku (Reasoning) vs Gemini 3 Flash Preview (Reasoning)
Anthropic vs Google — side-by-side benchmark comparison
| Claude 4.5 Haiku (Reasoning) | Gemini 3 Flash Preview (Reasoning) | |
|---|---|---|
| Intelligence Index | 37.1 | 46.4 |
| Coding Index | 32.6 | 42.6 |
| Math Index | 83.7 | 97.0 |
| Output speed (tok/s) | 142.2 | 192.1 |
| Blended price ($/1M) | $2.19 | $1.13 |
| Time to first token (s) | 10.48s | 6.35s |
| aime | — | — |
| aime 25 | 83.7% | 97.0% |
| artificial analysis coding index | 32.60 | 42.60 |
| artificial analysis intelligence index | 37.10 | 46.40 |
| artificial analysis math index | 83.70 | 97.00 |
| gpqa | 67.2% | 89.8% |
| hle | 9.7% | 34.7% |
| ifbench | 54.3% | 78.0% |
| lcr | 70.3% | 66.3% |
| livecodebench | 61.5% | 90.8% |
| math 500 | — | — |
| mmlu pro | 76.0% | 89.0% |
| scicode | 43.3% | 50.6% |
| tau2 | 54.7% | 80.4% |
| terminalbench hard | 27.3% | 38.6% |
Benchmark data from Artificial Analysis.