Across 14 shared benchmarks, Claude 3.5 Haiku scores higher on 1 and K-EXAONE 2.0 on 13. The widest gap is HLE, where K-EXAONE 2.0 scores 18.6 against 3.6.
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.