← All comparisons

Ministral 3 3B vs Hermes 4 - Llama-3.1 70B (Reasoning)

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

Ministral 3 3BHermes 4 - Llama-3.1 70B (Reasoning)
Intelligence Index11.216.0
Coding Index4.814.4
Math Index22.068.7
Output speed (tok/s)200.892.8
Blended price ($/1M)$0.10$0.20
Time to first token (s)0.34s0.64s
aime
aime 2522.0%68.7%
artificial analysis coding index4.8014.40
artificial analysis intelligence index11.2016.00
artificial analysis math index22.0068.70
gpqa35.8%69.9%
hle5.3%7.9%
ifbench26.8%31.3%
lcr11.7%6.7%
livecodebench24.7%65.3%
math 500
mmlu pro52.4%81.1%
scicode14.4%34.1%
tau224.9%22.5%
terminalbench hard0.0%4.5%

Benchmark data from Artificial Analysis.