← All comparisons
Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) vs Qwen3.6 27B (Reasoning)
NVIDIA vs Alibaba — side-by-side benchmark comparison
| Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) | Qwen3.6 27B (Reasoning) | |
|---|---|---|
| Intelligence Index | 15.0 | 45.8 |
| Coding Index | 13.1 | 36.5 |
| Math Index | 63.7 | — |
| Output speed (tok/s) | 52.3 | 57.4 |
| Blended price ($/1M) | $0.90 | $1.35 |
| Time to first token (s) | 0.76s | 1.46s |
| aime | 74.7% | — |
| aime 25 | 63.7% | — |
| artificial analysis coding index | 13.10 | 36.50 |
| artificial analysis intelligence index | 15.00 | 45.80 |
| artificial analysis math index | 63.70 | — |
| gpqa | 72.8% | 84.2% |
| hle | 8.1% | 21.6% |
| ifbench | 38.2% | 67.6% |
| lcr | 7.3% | 68.7% |
| livecodebench | 64.1% | — |
| math 500 | 95.2% | — |
| mmlu pro | 82.5% | — |
| scicode | 34.7% | 39.8% |
| tau2 | 11.4% | 94.2% |
| terminalbench hard | 2.3% | 34.8% |
Benchmark data from Artificial Analysis.