← All comparisons
Hermes 4 - Llama-3.1 70B (Reasoning) vs DeepSeek R1 Distill Qwen 14B
Nous Research vs DeepSeek — side-by-side benchmark comparison
| Hermes 4 - Llama-3.1 70B (Reasoning) | DeepSeek R1 Distill Qwen 14B | |
|---|---|---|
| Intelligence Index | 16.0 | 15.8 |
| Coding Index | 14.4 | — |
| Math Index | 68.7 | 55.7 |
| Output speed (tok/s) | 92.8 | 0.0 |
| Blended price ($/1M) | $0.20 | $0.00 |
| Time to first token (s) | 0.64s | 0.00s |
| aime | — | 66.7% |
| aime 25 | 68.7% | 55.7% |
| artificial analysis coding index | 14.40 | — |
| artificial analysis intelligence index | 16.00 | 15.80 |
| artificial analysis math index | 68.70 | 55.70 |
| gpqa | 69.9% | 48.4% |
| hle | 7.9% | 4.4% |
| ifbench | 31.3% | 22.1% |
| lcr | 6.7% | 7.0% |
| livecodebench | 65.3% | 37.6% |
| math 500 | — | 94.9% |
| mmlu pro | 81.1% | 74.0% |
| scicode | 34.1% | 23.9% |
| tau2 | 22.5% | — |
| terminalbench hard | 4.5% | — |
Benchmark data from Artificial Analysis.