← All comparisons
Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) vs Qwen3 235B A22B 2507 Instruct
NVIDIA vs Alibaba — side-by-side benchmark comparison
| Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) | Qwen3 235B A22B 2507 Instruct | |
|---|---|---|
| Intelligence Index | 15.0 | 25.0 |
| Coding Index | 13.1 | 22.1 |
| Math Index | 63.7 | 71.7 |
| Output speed (tok/s) | 52.3 | 57.0 |
| Blended price ($/1M) | $0.90 | $0.36 |
| Time to first token (s) | 0.76s | 1.34s |
| aime | 74.7% | 71.7% |
| aime 25 | 63.7% | 71.7% |
| artificial analysis coding index | 13.10 | 22.10 |
| artificial analysis intelligence index | 15.00 | 25.00 |
| artificial analysis math index | 63.70 | 71.70 |
| gpqa | 72.8% | 75.3% |
| hle | 8.1% | 10.6% |
| ifbench | 38.2% | 46.1% |
| lcr | 7.3% | 31.2% |
| livecodebench | 64.1% | 52.4% |
| math 500 | 95.2% | 98.0% |
| mmlu pro | 82.5% | 82.8% |
| scicode | 34.7% | 36.0% |
| tau2 | 11.4% | 33.3% |
| terminalbench hard | 2.3% | 15.2% |
Benchmark data from Artificial Analysis.