Benchmark Almanac

Hermes 4.3 36B

Hermes 4.3 36B ranks first on OpenBookQA.

Nous Research Released December 3rd, 2025 Family Hermes 4 The lab's model card Compare with others
13
Benchmarks with a score
1
Ranked first, out of five or more models
2
In the top three
Nous Research Dec 2025

Hermes 4.3 36B

Open weights Mid-size

131.2 #268 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
  • MMLU-Pro 80.7% #74 /167
  • AIME 2025 69.3% #113 /143
  • IFEval 77.9% #111 /135

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

BenchmarkScoreRankSource
OpenBookQA 96.6% #1 of 24 Lab model card
MMLU 87.7% #31 of 131 Lab model card
BIG-Bench Hard 86.4% #19 of 73 Lab model card
MMLU-Pro 80.7% #74 of 167 Lab model card
DROP 83.5% #19 of 41 Lab model card
MuSR 69.7% #3 of 6 Lab model card
IFEval 77.9% #111 of 135 Lab model card
SimpleQA 6.00% #57 of 60 Lab model card

Math and science

BenchmarkScoreRankSource
MATH-500 93.8% #46 of 70 Lab model card
GPQA Diamond 65.5% #204 of 302 Lab model card
AIME 2024 71.9% #60 of 88 Lab model card
AIME 2025 69.3% #113 of 143 Lab model card

Safety and security

BenchmarkScoreRankSource
RefusalBench (Nous Research) 74.6% #1 of 2 Lab model card

Its lab's numbers put it at 131.2, number 268 of 404

The Self-Reported Index fits one score to the 11 capability results Nous Research reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 128.0 to 134.4. How we compute it.

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.