Across 29 shared benchmarks, GPT-5.5 scores higher on 27 and Qwen3.5 122B A10B on 2. The widest gap is TauBench V3 - Banking, where GPT-5.5 scores 39 against 15.3. Qwen3.5 122B A10B is the cheaper of the two on tracked API pricing ($0.40 against $5.00 per million input tokens).
| Benchmark | GPT-5.5 | Qwen3.5 122B A10B |
|---|---|---|
| AA Agentic Index | 47.4 | 21.3 |
| AA Intelligence | 56.3 | 32.8 |
| AA-LCR | 79 | 70.3 |
| AA-Omniscience | 20.5 | -41.5 |
| arena_vision | 1293 | 1246 |
| Artificial Analysis Coding Index | 74.9 | 45.7 |
| browsecomp | 84.4 | 63.8 |
| coding_arena_elo | 1458 | 1358 |
| critpt | 27.1 | 0.9 |
| gdpval | 49.4 | 24.3 |
| GPQA Diamond | 93.6 | 86.6 |
| HLE | 52.2 | 47.5 |
| IFBench | 75.9 | 76.1 |
| MMMU-Pro | 83.2 | 76.9 |
| OmniScience Accuracy | 58 | 24.4 |
| OmniScience Non-Hallucination | 12 | 12.9 |
| OSWorld-Verified | 78.7 | 58 |
| scicode | 56.1 | 42 |
| SWE-bench Verified | 80.6 | 72 |
| TauBench V3 - Banking | 39 | 15.3 |
| Terminal-Bench 2.0 | 82.7 | 49.4 |
| Terminal-Bench 2.1 | 84.3 | 47.6 |
| Terminal-Bench Hard | 60.6 | 31.1 |
| vectara_answer_rate | 100 | 99.8 |
| vectara_avg_summary_length | 129.6 | 86.4 |
| vectara_factual_consistency | 90.7 | 88.8 |
| vectara_hallucination_rate ↓ | 9.3 | 11.2 |
| τ²-Bench Telecom (AA run) | 93.9 | 93.6 |
| τ³-Bench | 31.3 | 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 (direct, direct), otherwise the lowest tracked offer.