GPT-5.6 Sol vs Qwen3.5 397B A17B
GPT-5.6 Sol 
Qwen3.5 397B A17B Verdict
- GPT-5.6 Sol (high) leads on quality: 68.2 vs 58.7.
- GPT-5.6 Sol (high) is stronger in agents & tools, coding, composite, knowledge, long context, multimodal, reasoning.
- Qwen3.5 397B A17B is stronger in instruction following, human preference, maths.
- Qwen3.5 397B A17B is 21× cheaper ($0.387 vs $8.00 per 1M blended).
- Qwen3.5 397B A17B streams 1.3× faster (78 vs 59 tokens per second).
| Metric | GPT-5.6 Sol (high) | Qwen3.5 397B A17B |
|---|---|---|
| BenchLeader Index | 68.2 | 58.7 |
| Agents & tools score | 87.9 | 69.4 |
| Coding score | 72.4 | 51.7 |
| Composite score | 82.7 | 53.1 |
| Instruction following score | 67.8 | 76.1 |
| Knowledge score | 73.8 | 49.5 |
| Long context score | 67.4 | 65.2 |
| Multimodal score | 66.4 | 61.6 |
| Reasoning score | 63.3 | 63.0 |
| Human preference score | – | 63.6 |
| Maths score | – | 50.1 |
| Blended price $/M | $8.00 | $0.387 |
| Output speed | 59 tok/s | 78 tok/s |
| Time to first answer | 31.0 s | 42.7 s |
| Context window | 1M | 262k |
| GPQA Diamond | – | 85.9% |
| FrontierMath Tiers 1–3 | – | 29.5% |
| OTIS Mock AIME | – | 88.9% |
| SciCode | 56.9% | – |
| WeirdML | 88.8% | – |
| Epoch Capabilities Index | – | 147.0 |
| LMArena Text | – | 1441 |
| LMArena Hard Prompts | – | 1463 |
| LMArena Coding | – | 1491 |
| LMArena WebDev | – | 1399 |
| LMArena Vision | – | 1265 |
| AA Intelligence Index | 42.5 | 19.1 |
| IFBench | 69.2% | 78.8% |
| AA-LCR | 81.7% | 77.3% |
| MMMU-Pro | 81.8% | 77.3% |
| AA-Omniscience | 20.4 | -30.8 |
| Terminal-Bench Hard | 62.1% | 40.9% |
| GPQA Diamond (AA) | 92.8% | 89.3% |
| Humanity's Last Exam (AA) | 46.0% | 29.0% |
| SciCode (AA) | 57.8% | 44.8% |
| τ²-Bench Telecom (AA) | 83.3% | 95.6% |
| AIME 2026 | – | 94.2% |
| HMMT February 2026 | – | 87.9% |
| EnigmaEval | 37.1% | – |
| ARC-AGI-1 | 97.0% | – |
| ARC-AGI-2 | 85.4% | – |
| ARC-AGI-3 | 2.2% | – |
Data as of 2026-09-10. Best configuration of each model; every score links to its source on the model pages.
GPT-5.6 Sol vs Qwen3.5 397B A17B: questions
- Is GPT-5.6 Sol better than Qwen3.5 397B A17B?
- GPT-5.6 Sol (high) leads on quality: 68.2 vs 58.7. The BenchLeader Index combines every independent quality benchmark; GPT-5.6 Sol (high) is ahead overall as of 2026-09-10, but check the category scores for your use.
- Is GPT-5.6 Sol better than Qwen3.5 397B A17B for coding?
- GPT-5.6 Sol scores higher in coding (72 vs 52 on the category index, where 50 is average).
- Is GPT-5.6 Sol better than Qwen3.5 397B A17B for agentic tasks?
- GPT-5.6 Sol scores higher in agentic tasks (88 vs 69 on the category index, where 50 is average).
- Which is cheaper, GPT-5.6 Sol or Qwen3.5 397B A17B?
- Qwen3.5 397B A17B is cheaper: $0.387 against $8.00 per million tokens, blended at three input tokens per output token.
- Which is faster, GPT-5.6 Sol or Qwen3.5 397B A17B?
- Qwen3.5 397B A17B streams faster: 78 against 59 output tokens per second.
- Which has the larger context window?
- GPT-5.6 Sol accepts more context: 1M against 262k tokens.