← All comparisons
Hermes 4 - Llama-3.1 405B (Reasoning) vs GPT-4o (May '24)
Nous Research vs OpenAI — side-by-side benchmark comparison
| Hermes 4 - Llama-3.1 405B (Reasoning) | GPT-4o (May '24) | |
|---|---|---|
| Intelligence Index | 18.6 | 14.5 |
| Coding Index | 16.0 | 24.2 |
| Math Index | 69.7 | — |
| Output speed (tok/s) | 38.6 | 111.8 |
| Blended price ($/1M) | $1.50 | $7.50 |
| Time to first token (s) | 0.79s | 0.61s |
| aime | — | 11.0% |
| aime 25 | 69.7% | — |
| artificial analysis coding index | 16.00 | 24.20 |
| artificial analysis intelligence index | 18.60 | 14.50 |
| artificial analysis math index | 69.70 | — |
| gpqa | 72.7% | 52.6% |
| hle | 10.3% | 2.8% |
| ifbench | 32.7% | — |
| lcr | 20.7% | — |
| livecodebench | 68.6% | 33.4% |
| math 500 | — | 79.1% |
| mmlu pro | 82.9% | 74.0% |
| scicode | 25.2% | 30.9% |
| tau2 | 22.2% | — |
| terminalbench hard | 11.4% | — |
Benchmark data from Artificial Analysis.