GPT-5.6 Luna vs Qwen3 8
GPT-5.6 Luna 
Qwen3 8 Verdict
- Qwen3 8 (max) leads on quality: 66.5 vs 61.1.
- GPT-5.6 Luna (xhigh) is stronger in long context.
- Qwen3 8 (max) is stronger in agents & tools, coding, composite, human preference, knowledge, multimodal, reasoning.
- GPT-5.6 Luna (xhigh) is 5.9× cheaper ($0.450 vs $2.67 per 1M blended).
- GPT-5.6 Luna (xhigh) streams 3.2× faster (124 vs 39 tokens per second).
| Metric | GPT-5.6 Luna (xhigh) | Qwen3 8 (max) |
|---|---|---|
| BenchLeader Index | 61.1 | 66.5 |
| Agents & tools score | 54.1 | 68.3 |
| Coding score | 59.8 | 66.5 |
| Composite score | 73.1 | 74.3 |
| Human preference score | 64.8 | 68.2 |
| Knowledge score | 59.1 | 63.5 |
| Long context score | 67.3 | 66.6 |
| Maths score | 67.0 | 67.0 |
| Multimodal score | 61.6 | 67.4 |
| Reasoning score | 60.1 | 76.3 |
| Blended price $/M | $0.450 | $2.67 |
| Output speed | 124 tok/s | 39 tok/s |
| Time to first answer | 34.9 s | 55.1 s |
| Context window | 1.1M | 1M |
| SimpleBench | 46.8% | – |
| SciCode | 50.0% | 52.9% |
| ProofBench | – | 58.0% |
| Epoch Capabilities Index | – | 156.6 |
| LMArena Text | 1453 | 1481 |
| LMArena Hard Prompts | 1473 | 1503 |
| LMArena Coding | 1498 | 1522 |
| LMArena WebDev | 1519 | 1671 |
| LMArena Vision | 1259 | 1315 |
| LMArena Agent | -0.4 | 3.3 |
| LiveBench | – | 78.5% |
| LiveBench Reasoning | – | 88.2% |
| LiveBench Coding | – | 72.9% |
| LiveBench Agentic Coding | – | 64.7% |
| LiveBench Mathematics | – | 91.3% |
| LiveBench Data Analysis | – | 78.4% |
| LiveBench Language | – | 79.7% |
| LiveBench Instruction Following | – | 74.1% |
| AA Intelligence Index | 34.8 | 45.4 |
| AA-LCR | 81.7% | 80.3% |
| MMMU-Pro | 78.5% | 82.8% |
| AA-Omniscience | -10.8 | 12.0 |
| GPQA Diamond (AA) | 89.5% | 92.8% |
| Humanity's Last Exam (AA) | 37.0% | 43.1% |
| SciCode (AA) | 50.5% | 53.2% |
| LiveCodeBench | – | 87.8% |
| MMLU-Pro | – | 88.6% |
| IOI | – | 68.9% |
| LegalBench | – | 83.6% |
| CorpFin | – | 65.8% |
| TaxEval | – | 75.5% |
| Terminal-Bench 2.1 (Vals) | – | 67.4% |
| SWE-bench (Vals) | – | 85.6% |
| GPQA Diamond (Vals) | – | 93.7% |
| Vals Index | – | 51.8 |
| ARC-AGI-1 | 87.7% | – |
| ARC-AGI-2 | 47.6% | – |
| ARC-AGI-3 | 0.0% | – |
| CritPt | 20.6% | 20.0% |
| GDPval (AA) | 46.4% | 58.2% |
| τ²-Bench Banking (AA) | 28.7% | 51.3% |
| Code Migration | – | 24.0% |
| Excel Modeling Benchmark | – | 60.1% |
| Finance Agent v2 | – | 50.6% |
| Harvey's Legal Agent Benchmark | – | 10.4% |
| Legal Research Bench | – | 47.6% |
| MedCode | – | 40.7% |
| MedScribe | – | 85.0% |
| MMMU-Pro (Vals) | – | 88.0% |
| MortgageTax | – | 64.0% |
| MysteryMechanism | – | 23.9% |
| ProgramBench | – | 0.0% |
| Public Benefits Bench | – | 67.1% |
| SAGE | – | 51.3% |
| SkillsBench | – | 42.0% |
| Tax Agent Bench | – | 66.0% |
| Terminal-Bench 4.0 (Vals) | – | 24.8% |
| Terminal-Bench Science | – | 1.4% |
| Vals Multimodal Index | – | 65.4% |
| Vibe Code Bench 1-100 | – | 12.8% |
| Vibe Code Bench v1.1 | – | 64.7% |
| LMArena Maths | 1477 | 1498 |
| LMArena Creative Writing | 1411 | 1472 |
| LMArena Instruction Following | 1442 | 1477 |
| LMArena Multi-turn | 1458 | 1495 |
| LMArena Longer Queries | 1450 | 1492 |
| LMArena Document | 1462 | – |
| DeepSWE | 56.9% | – |
| CursorBench | 57.7% | – |
Data as of 2026-09-19. Best configuration of each model; every score links to its source on the model pages.
GPT-5.6 Luna vs Qwen3 8: questions
- Is GPT-5.6 Luna better than Qwen3 8?
- Qwen3 8 (max) leads on quality: 66.5 vs 61.1. The BenchLeader Index combines every independent quality benchmark; Qwen3 8 (max) is ahead overall as of 2026-09-19, but check the category scores for your use.
- Is GPT-5.6 Luna better than Qwen3 8 for coding?
- Qwen3 8 scores higher in coding (67 vs 60 on the category index, where 50 is average).
- Is GPT-5.6 Luna better than Qwen3 8 for agentic tasks?
- Qwen3 8 scores higher in agentic tasks (68 vs 54 on the category index, where 50 is average).
- Which is cheaper, GPT-5.6 Luna or Qwen3 8?
- GPT-5.6 Luna is cheaper: $0.450 against $2.67 per million tokens, blended at three input tokens per output token.
- Which is faster, GPT-5.6 Luna or Qwen3 8?
- GPT-5.6 Luna streams faster: 124 against 39 output tokens per second.
- Which has the larger context window?
- GPT-5.6 Luna accepts more context: 1.1M against 1M tokens.