Across 14 shared benchmarks, Celeris-1 scores higher on 0 and K-EXAONE 2.0 on 14. The widest gap is AA Agentic Index, where K-EXAONE 2.0 scores 19.7 against 2.4.
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.