← All comparisons

Hermes 4 - Llama-3.1 70B (Non-reasoning) vs Qwen3 VL 30B A3B Instruct

Nous Research vs Alibaba — side-by-side benchmark comparison

Hermes 4 - Llama-3.1 70B (Non-reasoning)Qwen3 VL 30B A3B Instruct
Intelligence Index12.616.0
Coding Index9.214.3
Math Index11.372.3
Output speed (tok/s)94.3123.5
Blended price ($/1M)$0.20$0.30
Time to first token (s)0.61s1.07s
aime
aime 2511.3%72.3%
artificial analysis coding index9.2014.30
artificial analysis intelligence index12.6016.00
artificial analysis math index11.3072.30
gpqa49.1%69.5%
hle3.6%6.4%
ifbench29.0%33.1%
lcr2.0%23.7%
livecodebench26.9%47.6%
math 500
mmlu pro66.4%76.4%
scicode27.7%30.8%
tau221.6%19.0%
terminalbench hard0.0%6.1%

Benchmark data from Artificial Analysis.