← All comparisons
Hermes 4 - Llama-3.1 70B (Reasoning) vs Claude Opus 4.6 (Adaptive Reasoning, Max Effort)
Nous Research vs Anthropic — side-by-side benchmark comparison
| Hermes 4 - Llama-3.1 70B (Reasoning) | Claude Opus 4.6 (Adaptive Reasoning, Max Effort) | |
|---|---|---|
| Intelligence Index | 16.0 | 52.9 |
| Coding Index | 14.4 | 48.1 |
| Math Index | 68.7 | — |
| Output speed (tok/s) | 92.8 | 54.8 |
| Blended price ($/1M) | $0.20 | $10.94 |
| Time to first token (s) | 0.64s | 11.69s |
| aime | — | — |
| aime 25 | 68.7% | — |
| artificial analysis coding index | 14.40 | 48.10 |
| artificial analysis intelligence index | 16.00 | 52.90 |
| artificial analysis math index | 68.70 | — |
| gpqa | 69.9% | 89.6% |
| hle | 7.9% | 36.7% |
| ifbench | 31.3% | 53.1% |
| lcr | 6.7% | 70.7% |
| livecodebench | 65.3% | — |
| math 500 | — | — |
| mmlu pro | 81.1% | — |
| scicode | 34.1% | 51.9% |
| tau2 | 22.5% | 92.1% |
| terminalbench hard | 4.5% | 46.2% |
Benchmark data from Artificial Analysis.