Claude Sonnet 4.5 vs Qwen3.5 4B
34 SHARED BENCHMARKSAcross 34 shared benchmarks, Claude Sonnet 4.5 scores higher on 30 and Qwen3.5 4B on 4. The widest gap is gdpval, where Claude Sonnet 4.5 scores 40.4 against 8.3. Qwen3.5 4B is the cheaper of the two on tracked API pricing ($0.03 against $3.00 per million input tokens).
ANTHROPICVSALIBABA34 SHARED30–4 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 (anthropic-official, direct), otherwise the lowest tracked offer.