← All comparisons
NVIDIA Nemotron Nano 12B v2 VL (Reasoning) vs Hermes 4 - Llama-3.1 70B (Non-reasoning)
NVIDIA vs Nous Research — side-by-side benchmark comparison
| NVIDIA Nemotron Nano 12B v2 VL (Reasoning) | Hermes 4 - Llama-3.1 70B (Non-reasoning) | |
|---|---|---|
| Intelligence Index | 14.9 | 12.6 |
| Coding Index | 11.7 | 9.2 |
| Math Index | 75.0 | 11.3 |
| Output speed (tok/s) | 0.0 | 94.3 |
| Blended price ($/1M) | $0.30 | $0.20 |
| Time to first token (s) | 0.00s | 0.61s |
| aime | — | — |
| aime 25 | 75.0% | 11.3% |
| artificial analysis coding index | 11.70 | 9.20 |
| artificial analysis intelligence index | 14.90 | 12.60 |
| artificial analysis math index | 75.00 | 11.30 |
| gpqa | 57.2% | 49.1% |
| hle | 5.3% | 3.6% |
| ifbench | 31.9% | 29.0% |
| lcr | 40.0% | 2.0% |
| livecodebench | 69.4% | 26.9% |
| math 500 | — | — |
| mmlu pro | 75.9% | 66.4% |
| scicode | 26.2% | 27.7% |
| tau2 | 21.3% | 21.6% |
| terminalbench hard | 4.5% | 0.0% |
Benchmark data from Artificial Analysis.