← All comparisons
DeepHermes 3 - Llama-3.1 8B Preview (Non-reasoning) vs GPT-5 (low)
Nous Research vs OpenAI — side-by-side benchmark comparison
| DeepHermes 3 - Llama-3.1 8B Preview (Non-reasoning) | GPT-5 (low) | |
|---|---|---|
| Intelligence Index | 7.6 | 39.2 |
| Coding Index | — | 30.7 |
| Math Index | — | 83.0 |
| Output speed (tok/s) | 0.0 | 88.9 |
| Blended price ($/1M) | $0.00 | $3.44 |
| Time to first token (s) | 0.00s | 8.32s |
| aime | 0.0% | 83.0% |
| aime 25 | — | 83.0% |
| artificial analysis coding index | — | 30.70 |
| artificial analysis intelligence index | 7.60 | 39.20 |
| artificial analysis math index | — | 83.00 |
| gpqa | 27.0% | 80.8% |
| hle | 4.3% | 18.4% |
| ifbench | — | 66.6% |
| lcr | — | 58.7% |
| livecodebench | 8.5% | 76.3% |
| math 500 | 21.8% | 98.7% |
| mmlu pro | 36.5% | 86.0% |
| scicode | 9.1% | 39.1% |
| tau2 | — | 84.2% |
| terminalbench hard | — | 26.5% |
Benchmark data from Artificial Analysis.