← All comparisons

Hermes 4 - Llama-3.1 405B (Non-reasoning) vs Cogito v2.1 (Reasoning)

Nous Research vs Deep Cogito — side-by-side benchmark comparison

Hermes 4 - Llama-3.1 405B (Non-reasoning)Cogito v2.1 (Reasoning)
Intelligence Index17.6
Coding Index18.124.8
Math Index15.372.7
Output speed (tok/s)40.880.7
Blended price ($/1M)$1.50$1.25
Time to first token (s)0.73s0.51s
aime
aime 2515.3%72.7%
artificial analysis coding index18.1024.80
artificial analysis intelligence index17.60
artificial analysis math index15.3072.70
gpqa53.6%76.8%
hle4.2%11.0%
ifbench34.8%46.3%
lcr20.0%21.7%
livecodebench54.6%68.8%
math 500
mmlu pro72.9%84.9%
scicode34.6%41.0%
tau226.6%
terminalbench hard9.8%16.7%

Benchmark data from Artificial Analysis.