Across 23 shared benchmarks, Claude Sonnet 5 scores higher on 22 and Qwen3.5 122B A10B on 1. The widest gap is OmniScience Non-Hallucination, where Claude Sonnet 5 scores 60.6 against 12.9. Qwen3.5 122B A10B is the cheaper of the two on tracked API pricing ($0.40 against $2.00 per million input tokens).
| Benchmark | Claude Sonnet 5 | Qwen3.5 122B A10B |
|---|---|---|
| AA Agentic Index | 49.7 | 21.3 |
| AA Intelligence | 55.3 | 32.8 |
| AA-LCR | 77 | 70.3 |
| AA-Omniscience | 16.4 | -41.5 |
| arena_vision | 1281 | 1246 |
| Artificial Analysis Coding Index | 71.5 | 45.7 |
| browsecomp | 84.7 | 63.8 |
| coding_arena_elo | 1539 | 1358 |
| critpt | 16.9 | 0.9 |
| gdpval | 54.6 | 24.3 |
| GPQA Diamond | 91.1 | 86.6 |
| HLE | 57.4 | 47.5 |
| LVBench | 68.5 | 74.4 |
| MMMU-Pro | 77.3 | 76.9 |
| OmniScience Accuracy | 40 | 24.4 |
| OmniScience Non-Hallucination | 60.6 | 12.9 |
| OSWorld-Verified | 81.2 | 58 |
| scicode | 53.6 | 42 |
| SWE-bench Verified | 85.2 | 72 |
| TauBench V3 - Banking | 37.3 | 15.3 |
| Terminal-Bench 2.0 | 80.4 | 49.4 |
| Terminal-Bench 2.1 | 80.5 | 47.6 |
| τ³-Bench | 28.2 | 13.6 |
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.