← All comparisons
Hy3-preview (Reasoning) vs Llama 3.3 Nemotron Super 49B v1 (Reasoning)
Tencent vs NVIDIA — side-by-side benchmark comparison
| Hy3-preview (Reasoning) | Llama 3.3 Nemotron Super 49B v1 (Reasoning) | |
|---|---|---|
| Intelligence Index | 41.9 | 18.5 |
| Coding Index | 36.5 | 9.4 |
| Math Index | — | 54.7 |
| Output speed (tok/s) | 107.7 | 0.0 |
| Blended price ($/1M) | $0.20 | $0.00 |
| Time to first token (s) | 2.46s | 0.00s |
| aime | — | 58.3% |
| aime 25 | — | 54.7% |
| artificial analysis coding index | 36.50 | 9.40 |
| artificial analysis intelligence index | 41.90 | 18.50 |
| artificial analysis math index | — | 54.70 |
| gpqa | 86.7% | 64.3% |
| hle | 25.5% | 6.5% |
| ifbench | 63.1% | 38.1% |
| lcr | 54.7% | 17.0% |
| livecodebench | — | 27.7% |
| math 500 | — | 95.9% |
| mmlu pro | — | 78.5% |
| scicode | 41.2% | 28.2% |
| tau2 | 92.7% | 26.9% |
| terminalbench hard | 34.1% | 0.0% |
Benchmark data from Artificial Analysis.