Across 17 shared benchmarks, Kimi K3 scores higher on 16 and Qwen3 VL 235B A22B Reasoning on 1. The widest gap is ZeroBench, where Qwen3 VL 235B A22B Reasoning scores 4 against 0.4. Qwen3 VL 235B A22B Reasoning is the cheaper of the two on tracked API pricing ($0.40 against $3.00 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 (direct, alibaba-official), otherwise the lowest tracked offer.