GPT-5.5 Pro vs Qwen3.7 Plus
GPT-5.5 Pro 
Qwen3.7 Plus Verdict
- GPT-5.5 Pro (xhigh) leads on quality: 67.1 vs 59.9.
- GPT-5.5 Pro (xhigh) is stronger in maths, reasoning.
- Qwen3.7 Plus is stronger in agents & tools, coding, composite, human preference, instruction following, knowledge, long context, multimodal.
- Qwen3.7 Plus is 60× cheaper ($1.13 vs $67.50 per 1M blended).
- Qwen3.7 Plus streams 16.6× faster (66 vs 4 tokens per second).
| Metric | GPT-5.5 Pro (xhigh) | Qwen3.7 Plus |
|---|---|---|
| BenchLeader Index | 67.1 | 59.9 |
| Maths score | 74.5 | 65.6 |
| Reasoning score | 80.1 | 62.4 |
| Agents & tools score | – | 45.2 |
| Coding score | – | 46.7 |
| Composite score | – | 61.7 |
| Human preference score | – | 65.2 |
| Instruction following score | – | 75.6 |
| Knowledge score | – | 64.7 |
| Long context score | – | 62.8 |
| Multimodal score | – | 63.9 |
| Blended price $/M | $67.50 | $1.13 |
| Output speed | 4 tok/s | 66 tok/s |
| Time to first answer | 3.5 s | 32.4 s |
| Context window | 1.1M | 1M |
| GPQA Diamond | – | 87.9% |
| FrontierMath Tiers 1–3 | 87.7% | – |
| FrontierMath Tier 4 | 78.0% | – |
| OTIS Mock AIME | – | 93.3% |
| OSWorld-Verified 2.0 | – | 2.8% |
| SciCode | – | 45.5% |
| FrontierCode | – | 10.2% |
| Epoch Capabilities Index | – | 147.4 |
| LMArena Text | – | 1456 |
| LMArena Hard Prompts | – | 1473 |
| LMArena Coding | – | 1501 |
| LMArena Vision | – | 1278 |
| LMArena Agent | – | -6.1 |
| AA Intelligence Index | – | 25.8 |
| IFBench | – | 78.0% |
| AA-LCR | – | 73.0% |
| MMMU-Pro | – | 80.5% |
| AA-Omniscience | – | 1.1 |
| Terminal-Bench Hard | – | 47.0% |
| GPQA Diamond (AA) | – | 90.0% |
| Humanity's Last Exam (AA) | – | 35.6% |
| SciCode (AA) | – | 46.1% |
| τ²-Bench Telecom (AA) | – | 93.0% |
| Terminal-Bench 2.1 (Vals) | – | 52.8% |
| Vals Index | – | 38.6 |
| ARC-AGI-1 | 95.0% | – |
| ARC-AGI-2 | 84.2% | – |
| CritPt | 30.6% | 9.1% |
| GDPval (AA) | – | 19.3% |
| τ²-Bench Banking (AA) | – | 17.5% |
| APEX-Agents (AA) | – | 22.4% |
| Code Migration | – | 12.9% |
| CyberBench | – | 37.5% |
| Excel Modeling Benchmark | – | 49.4% |
| Finance Agent v2 | – | 38.2% |
| Harvey's Legal Agent Benchmark | – | 0.0% |
| Legal Research Bench | – | 16.4% |
| MortgageTax | – | 66.2% |
| SAGE | – | 39.3% |
| SkillsBench | – | 54.3% |
| Vals Multimodal Index | – | 53.9% |
| Vibe Code Bench v1.1 | – | 46.4% |
| LMArena Maths | – | 1466 |
| LMArena Creative Writing | – | 1432 |
| LMArena Instruction Following | – | 1447 |
| LMArena Multi-turn | – | 1465 |
| LMArena Longer Queries | – | 1465 |
| LMArena Document | – | 1453 |
| Chess Puzzles | – | 24.0% |
| Mystery Game Puzzles | – | 17.0% |
| LMCA | 53.9% | 37.6% |
| DTBench | 96.0% | 84.0% |
Data as of 2026-09-19. Best configuration of each model; every score links to its source on the model pages.
GPT-5.5 Pro vs Qwen3.7 Plus: questions
- Is GPT-5.5 Pro better than Qwen3.7 Plus?
- GPT-5.5 Pro (xhigh) leads on quality: 67.1 vs 59.9. The BenchLeader Index combines every independent quality benchmark; GPT-5.5 Pro (xhigh) is ahead overall as of 2026-09-19, but check the category scores for your use.
- Which is cheaper, GPT-5.5 Pro or Qwen3.7 Plus?
- Qwen3.7 Plus is cheaper: $1.13 against $67.50 per million tokens, blended at three input tokens per output token.
- Which is faster, GPT-5.5 Pro or Qwen3.7 Plus?
- Qwen3.7 Plus streams faster: 66 against 4 output tokens per second.
- Which has the larger context window?
- GPT-5.5 Pro accepts more context: 1.1M against 1M tokens.