Across 22 shared benchmarks, GPT-4o mini scores higher on 0 and Kimi K2 Instruct on 22. The widest gap is SWE-bench Verified, where Kimi K2 Instruct scores 69.2 against 8.7. GPT-4o mini is the cheaper of the two on tracked API pricing ($0.15 against $0.60 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, direct), otherwise the lowest tracked offer.