← All comparisons

Granite 4.0 H 1B vs Hermes 3 - Llama-3.1 70B

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

Granite 4.0 H 1BHermes 3 - Llama-3.1 70B
Intelligence Index8.010.6
Coding Index2.7
Math Index6.3
Output speed (tok/s)0.033.2
Blended price ($/1M)$0.00$0.30
Time to first token (s)0.00s0.38s
aime2.3%
aime 256.3%
artificial analysis coding index2.70
artificial analysis intelligence index8.0010.60
artificial analysis math index6.30
gpqa26.3%40.1%
hle5.0%4.1%
ifbench26.2%
lcr6.3%
livecodebench11.5%18.8%
math 50053.8%
mmlu pro27.7%57.1%
scicode8.2%23.1%
tau219.6%
terminalbench hard0.0%

Benchmark data from Artificial Analysis.