← All comparisons
Claude 4.5 Haiku (Reasoning) vs Gemini 2.5 Flash-Lite (Reasoning)
Anthropic vs Google — side-by-side benchmark comparison
| Claude 4.5 Haiku (Reasoning) | Gemini 2.5 Flash-Lite (Reasoning) | |
|---|---|---|
| Intelligence Index | 37.1 | 17.6 |
| Coding Index | 32.6 | 9.5 |
| Math Index | 83.7 | 53.3 |
| Output speed (tok/s) | 142.2 | 275.9 |
| Blended price ($/1M) | $2.19 | $0.17 |
| Time to first token (s) | 10.48s | 17.31s |
| aime | — | 70.3% |
| aime 25 | 83.7% | 53.3% |
| artificial analysis coding index | 32.60 | 9.50 |
| artificial analysis intelligence index | 37.10 | 17.60 |
| artificial analysis math index | 83.70 | 53.30 |
| gpqa | 67.2% | 62.5% |
| hle | 9.7% | 6.4% |
| ifbench | 54.3% | 49.9% |
| lcr | 70.3% | 51.3% |
| livecodebench | 61.5% | 59.3% |
| math 500 | — | 96.9% |
| mmlu pro | 76.0% | 75.9% |
| scicode | 43.3% | 19.3% |
| tau2 | 54.7% | 18.4% |
| terminalbench hard | 27.3% | 4.5% |
Benchmark data from Artificial Analysis.