GPT-4o vs Qwen3.5 4B
35 SHARED BENCHMARKSAcross 35 shared benchmarks, GPT-4o scores higher on 8 and Qwen3.5 4B on 26, with 1 level. The widest gap is OmniScience Non-Hallucination, where GPT-4o scores 62.1 against 13.4. Qwen3.5 4B is the cheaper of the two on tracked API pricing ($0.03 against $5.00 per million input tokens).
OPENAIVSALIBABA35 SHARED8–26 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.