Across 33 shared benchmarks, Qwen3.5 122B A10B scores higher on 3 and Qwen3.7 Max on 30. The widest gap is OmniScience Non-Hallucination, where Qwen3.7 Max scores 74.4 against 12.9. Qwen3.5 122B A10B is the cheaper of the two on tracked API pricing ($0.40 against $2.50 per million input tokens).
| Benchmark | Qwen3.5 122B A10B | Qwen3.7 Max |
|---|---|---|
| AA Agentic Index | 21.3 | 30.9 |
| AA Intelligence | 32.8 | 46.7 |
| AA-LCR | 70.3 | 74.7 |
| AA-Omniscience | -41.5 | 13.5 |
| Artificial Analysis Coding Index | 45.7 | 66 |
| bfcl_v4 | 72.2 | 75 |
| coding_arena_elo | 1358 | 1517 |
| critpt | 0.9 | 13.4 |
| gdpval | 24.3 | 38.4 |
| GPQA Diamond | 86.6 | 92.4 |
| HLE | 47.5 | 41.4 |
| IFBench | 76.1 | 80.5 |
| ifeval | 93.4 | 94.3 |
| include | 82.8 | 86.2 |
| LiveCodeBench v6 | 78.9 | 91.6 |
| longbench_v2 | 60.2 | 65.3 |
| mmlu_prox | 82.2 | 87 |
| mmlu_redux | 94 | 95 |
| MMLU-Pro | 86.7 | 89.6 |
| mmmlu | 86.7 | 90.3 |
| OmniScience Accuracy | 24.4 | 31.1 |
| OmniScience Non-Hallucination | 12.9 | 74.4 |
| scicode | 42 | 53.5 |
| supergpqa | 67.1 | 73.6 |
| SWE-bench Verified | 72 | 80.4 |
| TauBench V3 - Banking | 15.3 | 11.8 |
| Terminal-Bench 2.0 | 49.4 | 69.7 |
| Terminal-Bench 2.1 | 47.6 | 75 |
| Terminal-Bench Hard | 31.1 | 50.8 |
| VITA-Bench | 33.6 | 47.9 |
| WideSearch | 60.5 | 75.2 |
| τ²-Bench Telecom (AA run) | 93.6 | 94.7 |
| τ³-Bench | 13.6 | 10.9 |
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.