← All comparisons
Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning) vs Qwen3 4B 2507 (Reasoning)
NVIDIA vs Alibaba — side-by-side benchmark comparison
| Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning) | Qwen3 4B 2507 (Reasoning) | |
|---|---|---|
| Intelligence Index | 14.4 | 18.2 |
| Coding Index | — | 9.5 |
| Math Index | 50.0 | 82.7 |
| Output speed (tok/s) | 0.0 | 0.0 |
| Blended price ($/1M) | $0.00 | $0.00 |
| Time to first token (s) | 0.00s | 0.00s |
| aime | 70.7% | — |
| aime 25 | 50.0% | 82.7% |
| artificial analysis coding index | — | 9.50 |
| artificial analysis intelligence index | 14.40 | 18.20 |
| artificial analysis math index | 50.00 | 82.70 |
| gpqa | 40.8% | 66.7% |
| hle | 5.1% | 5.9% |
| ifbench | 25.5% | 49.8% |
| lcr | 0.0% | 37.7% |
| livecodebench | 49.3% | 64.1% |
| math 500 | 94.7% | — |
| mmlu pro | 55.6% | 74.3% |
| scicode | 10.1% | 25.6% |
| tau2 | 11.7% | 25.4% |
| terminalbench hard | — | 1.5% |
Benchmark data from Artificial Analysis.