Across 14 shared benchmarks, Claude Fable 5.1 scores higher on 13 and K2 Horizon 375B A23B on 1. The widest gap is critpt, where Claude Fable 5.1 scores 31.1 against 4.6.
Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.