← All comparisons
Llama Nemotron Super 49B v1.5 (Non-reasoning) vs GPT-5 (minimal)
NVIDIA vs OpenAI — side-by-side benchmark comparison
| Llama Nemotron Super 49B v1.5 (Non-reasoning) | GPT-5 (minimal) | |
|---|---|---|
| Intelligence Index | 14.6 | 23.9 |
| Coding Index | 10.5 | 25.0 |
| Math Index | 8.0 | 31.7 |
| Output speed (tok/s) | 48.9 | 76.8 |
| Blended price ($/1M) | $0.17 | $3.44 |
| Time to first token (s) | 0.28s | 0.84s |
| aime | 13.7% | 36.7% |
| aime 25 | 8.0% | 31.7% |
| artificial analysis coding index | 10.50 | 25.00 |
| artificial analysis intelligence index | 14.60 | 23.90 |
| artificial analysis math index | 8.00 | 31.70 |
| gpqa | 48.1% | 67.3% |
| hle | 4.3% | 5.4% |
| ifbench | 32.9% | 45.6% |
| lcr | 22.0% | 25.0% |
| livecodebench | 29.0% | 55.8% |
| math 500 | 77.0% | 86.1% |
| mmlu pro | 69.2% | 80.6% |
| scicode | 23.8% | 38.8% |
| tau2 | 25.1% | 67.0% |
| terminalbench hard | 3.8% | 18.2% |
Benchmark data from Artificial Analysis.