← All comparisons
Claude 4.5 Haiku (Reasoning) vs Hermes 4 - Llama-3.1 70B (Non-reasoning)
Anthropic vs Nous Research — side-by-side benchmark comparison
| Claude 4.5 Haiku (Reasoning) | Hermes 4 - Llama-3.1 70B (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 37.1 | 12.6 |
| Coding Index | 32.6 | 9.2 |
| Math Index | 83.7 | 11.3 |
| Output speed (tok/s) | 142.2 | 94.3 |
| Blended price ($/1M) | $2.19 | $0.20 |
| Time to first token (s) | 10.48s | 0.61s |
| aime | — | — |
| aime 25 | 83.7% | 11.3% |
| artificial analysis coding index | 32.60 | 9.20 |
| artificial analysis intelligence index | 37.10 | 12.60 |
| artificial analysis math index | 83.70 | 11.30 |
| gpqa | 67.2% | 49.1% |
| hle | 9.7% | 3.6% |
| ifbench | 54.3% | 29.0% |
| lcr | 70.3% | 2.0% |
| livecodebench | 61.5% | 26.9% |
| math 500 | — | — |
| mmlu pro | 76.0% | 66.4% |
| scicode | 43.3% | 27.7% |
| tau2 | 54.7% | 21.6% |
| terminalbench hard | 27.3% | 0.0% |
Benchmark data from Artificial Analysis.