← All comparisons
Hermes 3 - Llama-3.1 70B vs Qwen3 VL 4B (Reasoning)
Nous Research vs Alibaba — side-by-side benchmark comparison
| Hermes 3 - Llama-3.1 70B | Qwen3 VL 4B (Reasoning) | |
|---|---|---|
| Intelligence Index | 10.6 | 13.7 |
| Coding Index | — | 6.7 |
| Math Index | — | 25.7 |
| Output speed (tok/s) | 33.2 | 0.0 |
| Blended price ($/1M) | $0.30 | $0.00 |
| Time to first token (s) | 0.38s | 0.00s |
| aime | 2.3% | — |
| aime 25 | — | 25.7% |
| artificial analysis coding index | — | 6.70 |
| artificial analysis intelligence index | 10.60 | 13.70 |
| artificial analysis math index | — | 25.70 |
| gpqa | 40.1% | 49.4% |
| hle | 4.1% | 4.4% |
| ifbench | — | 36.6% |
| lcr | — | 21.3% |
| livecodebench | 18.8% | 32.0% |
| math 500 | 53.8% | — |
| mmlu pro | 57.1% | 70.0% |
| scicode | 23.1% | 17.1% |
| tau2 | — | 15.5% |
| terminalbench hard | — | 1.5% |
Benchmark data from Artificial Analysis.