Across 25 shared benchmarks, Muse Spark 1.3 scores higher on 11 and Claude Opus 5 on 14. The widest gap is OmniScience Non-Hallucination, where Muse Spark 1.3 scores 68.5 against 40.5. Muse Spark 1.3 is the cheaper of the two on tracked API pricing ($1.25 against $5.00 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 (meta-official, direct), otherwise the lowest tracked offer.