← All comparisons

Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) vs Qwen2.5 Turbo

NVIDIA vs Alibaba — side-by-side benchmark comparison

Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)Qwen2.5 Turbo
Intelligence Index15.012.0
Coding Index13.1
Math Index63.7
Output speed (tok/s)52.366.5
Blended price ($/1M)$0.90$0.09
Time to first token (s)0.76s1.16s
aime74.7%12.0%
aime 2563.7%
artificial analysis coding index13.10
artificial analysis intelligence index15.0012.00
artificial analysis math index63.70
gpqa72.8%41.0%
hle8.1%4.2%
ifbench38.2%
lcr7.3%
livecodebench64.1%16.3%
math 50095.2%80.5%
mmlu pro82.5%63.3%
scicode34.7%15.3%
tau211.4%
terminalbench hard2.3%

Benchmark data from Artificial Analysis.