← All comparisons

Hermes 4 - Llama-3.1 70B (Reasoning) vs Kimi K2.5 (Reasoning)

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

Hermes 4 - Llama-3.1 70B (Reasoning)Kimi K2.5 (Reasoning)
Intelligence Index16.046.8
Coding Index14.439.6
Math Index68.7
Output speed (tok/s)92.833.7
Blended price ($/1M)$0.20$1.19
Time to first token (s)0.64s1.29s
aime
aime 2568.7%
artificial analysis coding index14.4039.60
artificial analysis intelligence index16.0046.80
artificial analysis math index68.70
gpqa69.9%87.9%
hle7.9%29.4%
ifbench31.3%70.2%
lcr6.7%65.3%
livecodebench65.3%
math 500
mmlu pro81.1%
scicode34.1%49.0%
tau222.5%95.9%
terminalbench hard4.5%34.8%

Benchmark data from Artificial Analysis.