← All benchmarks
Hermes 4 - Llama-3.1 70B (Reasoning)
Nous Research · Released 2025-08-27 · Independent benchmark scores from Artificial Analysis
Intelligence Index
10.0
Output speed
95.7
tokens/sec
Blended price
$0.20
per 1M tokens
Time to first token
0.55s
Model metadata
| Model ID | 6ba9e8eb-8124-436d-842f-dbe36df80c27 |
|---|---|
| Slug | hermes-4-llama-3-1-70b-reasoning |
| Release date | 2025-08-27 |
| Provider | Nous Research |
| Provider slug | nous-research |
| Output speed | 95.7 tok/s |
| Time to first token | 0.55s |
| Time to first answer | 21.45s |
| Blended price ($/1M) | $0.20 |
| Input price ($/1M) | $0.13 |
| Output price ($/1M) | $0.40 |
All evaluation scores (17)
| aime | — |
|---|---|
| aime 25 | 68.7% |
| artificial analysis coding index | — |
| artificial analysis intelligence index | 10.00 |
| artificial analysis math index | 68.70 |
| gpqa | 69.9% |
| hle | 7.9% |
| ifbench | 31.3% |
| lcr | 6.7% |
| livecodebench | 65.3% |
| math 500 | — |
| mmlu pro | 81.1% |
| scicode | 34.1% |
| tau banking | — |
| tau2 | 22.5% |
| terminalbench hard | 4.5% |
| terminalbench v2 1 | — |
Source: artificialanalysis.ai/models/hermes-4-llama-3-1-70b-reasoning
Benchmark data from Artificial Analysis.