← All comparisons

Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) vs GPT-5 mini (medium)

NVIDIA vs OpenAI — side-by-side benchmark comparison

Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)GPT-5 mini (medium)
Intelligence Index15.038.9
Coding Index13.132.8
Math Index63.785.0
Output speed (tok/s)52.392.4
Blended price ($/1M)$0.90$0.69
Time to first token (s)0.76s16.52s
aime74.7%
aime 2563.7%85.0%
artificial analysis coding index13.1032.80
artificial analysis intelligence index15.0038.90
artificial analysis math index63.7085.00
gpqa72.8%80.3%
hle8.1%14.6%
ifbench38.2%71.2%
lcr7.3%66.0%
livecodebench64.1%69.2%
math 50095.2%
mmlu pro82.5%82.8%
scicode34.7%41.0%
tau211.4%71.1%
terminalbench hard2.3%28.8%

Benchmark data from Artificial Analysis.