← All comparisons

Granite 4.0 H 1B vs Hermes 4 - Llama-3.1 70B (Non-reasoning)

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

Granite 4.0 H 1BHermes 4 - Llama-3.1 70B (Non-reasoning)
Intelligence Index8.012.6
Coding Index2.79.2
Math Index6.311.3
Output speed (tok/s)0.094.3
Blended price ($/1M)$0.00$0.20
Time to first token (s)0.00s0.61s
aime
aime 256.3%11.3%
artificial analysis coding index2.709.20
artificial analysis intelligence index8.0012.60
artificial analysis math index6.3011.30
gpqa26.3%49.1%
hle5.0%3.6%
ifbench26.2%29.0%
lcr6.3%2.0%
livecodebench11.5%26.9%
math 500
mmlu pro27.7%66.4%
scicode8.2%27.7%
tau219.6%21.6%
terminalbench hard0.0%0.0%

Benchmark data from Artificial Analysis.