Across 14 shared benchmarks, Claude 3.5 Haiku scores higher on 4 and MiniCPM5-2B on 10. The widest gap is AA Agentic Index, where MiniCPM5-2B scores 9 against 1.9.
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.