GPT-5.2 vs Qwen3.5 4B
29 SHARED BENCHMARKSAcross 29 shared benchmarks, GPT-5.2 scores higher on 27 and Qwen3.5 4B on 2. The widest gap is gdpval, where GPT-5.2 scores 48.3 against 8.3. Qwen3.5 4B is the cheaper of the two on tracked API pricing ($0.03 against $1.75 per million input tokens).
OPENAIVSALIBABA29 SHARED27–2 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, direct), otherwise the lowest tracked offer.