← All comparisons

Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) vs Qwen3 VL 30B A3B (Reasoning)

NVIDIA vs Alibaba — side-by-side benchmark comparison

Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)Qwen3 VL 30B A3B (Reasoning)
Intelligence Index15.019.7
Coding Index13.113.1
Math Index63.782.3
Output speed (tok/s)52.3123.6
Blended price ($/1M)$0.90$0.34
Time to first token (s)0.76s1.09s
aime74.7%
aime 2563.7%82.3%
artificial analysis coding index13.1013.10
artificial analysis intelligence index15.0019.70
artificial analysis math index63.7082.30
gpqa72.8%72.0%
hle8.1%8.7%
ifbench38.2%45.1%
lcr7.3%40.7%
livecodebench64.1%69.7%
math 50095.2%
mmlu pro82.5%80.7%
scicode34.7%28.8%
tau211.4%19.9%
terminalbench hard2.3%5.3%

Benchmark data from Artificial Analysis.