← All comparisons
Llama 3.3 Nemotron Super 49B v1 (Non-reasoning) vs Qwen3 VL 8B (Reasoning)
NVIDIA vs Alibaba — side-by-side benchmark comparison
| Llama 3.3 Nemotron Super 49B v1 (Non-reasoning) | Qwen3 VL 8B (Reasoning) | |
|---|---|---|
| Intelligence Index | 14.3 | 16.7 |
| Coding Index | 7.6 | 9.8 |
| Math Index | 7.7 | 30.7 |
| Output speed (tok/s) | 0.0 | 133.6 |
| Blended price ($/1M) | $0.00 | $0.66 |
| Time to first token (s) | 0.00s | 1.13s |
| aime | 19.3% | — |
| aime 25 | 7.7% | 30.7% |
| artificial analysis coding index | 7.60 | 9.80 |
| artificial analysis intelligence index | 14.30 | 16.70 |
| artificial analysis math index | 7.70 | 30.70 |
| gpqa | 51.7% | 57.9% |
| hle | 3.5% | 3.3% |
| ifbench | 39.5% | 39.9% |
| lcr | 11.3% | 31.0% |
| livecodebench | 28.0% | 35.3% |
| math 500 | 77.5% | — |
| mmlu pro | 69.8% | 74.9% |
| scicode | 22.9% | 21.9% |
| tau2 | — | 22.5% |
| terminalbench hard | 0.0% | 3.8% |
Benchmark data from Artificial Analysis.