Across 15 shared benchmarks, Agnes 2.5 Pro Beta scores higher on 15 and Claude 3.5 Haiku on 0. The widest gap is Terminal-Bench 2.1, where Agnes 2.5 Pro Beta scores 69.7 against 10.1. Agnes 2.5 Pro Beta is the cheaper of the two on tracked API pricing ($0.10 against $0.80 per million input tokens).
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. Quoted rates are the price-setter row we currently track for each model — its direct or vendor-official listing where one exists (direct, anthropic-official), otherwise the lowest tracked offer.