Benchmarks

FrontierMath (Tier 4)

Mathunit % · normalized over [0, 88]

FrontierMath hardest tier — IMO/research-adjacent difficulty.

#ModelSourceScoreNormalized

Score distribution

11 tracked results across the normalization window

088

Score vs. parameters

Open-weights models, log-x params

No open-weights models with disclosed parameter counts have a score here yet.