← All comparisons
Claude 4.5 Haiku (Reasoning) vs Kimi K2 Thinking
Anthropic vs Kimi — side-by-side benchmark comparison
| Claude 4.5 Haiku (Reasoning) | Kimi K2 Thinking | |
|---|---|---|
| Intelligence Index | 37.1 | 40.9 |
| Coding Index | 32.6 | 34.8 |
| Math Index | 83.7 | 94.7 |
| Output speed (tok/s) | 142.2 | 130.3 |
| Blended price ($/1M) | $2.19 | $1.07 |
| Time to first token (s) | 10.48s | 0.84s |
| aime | — | — |
| aime 25 | 83.7% | 94.7% |
| artificial analysis coding index | 32.60 | 34.80 |
| artificial analysis intelligence index | 37.10 | 40.90 |
| artificial analysis math index | 83.70 | 94.70 |
| gpqa | 67.2% | 83.8% |
| hle | 9.7% | 22.3% |
| ifbench | 54.3% | 68.1% |
| lcr | 70.3% | 66.3% |
| livecodebench | 61.5% | 85.3% |
| math 500 | — | — |
| mmlu pro | 76.0% | 84.8% |
| scicode | 43.3% | 42.4% |
| tau2 | 54.7% | 93.0% |
| terminalbench hard | 27.3% | 31.1% |
Benchmark data from Artificial Analysis.