Qwen3.5 4B vs Qwen3.6 27B
35 SHARED BENCHMARKSAcross 35 shared benchmarks, Qwen3.5 4B scores higher on 2 and Qwen3.6 27B on 33. The widest gap is OmniScience Non-Hallucination, where Qwen3.6 27B scores 50.7 against 13.4. Qwen3.5 4B is the cheaper of the two on tracked API pricing ($0.03 against $0.60 per million input tokens).
ALIBABAVSALIBABA35 SHARED2–33 HEAD-TO-HEAD
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.