Model comparison
GPT-5.6 Sol vs Qwen3.8 Max
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 56.8 on the Noometry Index. Qwen3.8 Max costs 2.7× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Last verified . 39 shared benchmarks.
Alibaba (Qwen)Summary
- They share 39 benchmarks with published results for both. GPT-5.6 Sol scores higher in 8 categories and Qwen3.8 Max in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.6 Sol leads 74.8 to 54.4.
- The biggest single-benchmark swing is Furniture Assembly: 56.7% for GPT-5.6 Sol and 20% for Qwen3.8 Max.
- Qwen3.8 Max is cheaper at $2 / $6 per million input/output tokens, against $4 / $20 for GPT-5.6 Sol.
- GPT-5.6 Sol accepts more context: 1.05M tokens versus 1M.
Side by side
| GPT-5.6 Sol | Qwen3.8 Max | |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 65.0 | 56.8 |
| Released | 2026-07-09 | 2026-08-02 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 1M |
| Max output | 128K | 131K |
| Input $ / M tokens | $4 | $2 |
| Output $ / M tokens | $20 | $6 |
| Results tracked | 65 | 39 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Sol leads
GPT-5.6 Sol: 65.1 (#7), Qwen3.8 Max: 53.5 (#29)
| Benchmark | GPT-5.6 Sol | Qwen3.8 Max |
|---|---|---|
| DeepSWE | 72.7% | 57.5% |
| LMArena WebDev | 1618 | 1674 |
| FrontierSWE | 32.2% | 17.8% |
| SciCode | 57.1% | 53.2% |
| LMArena Coding | 1498 | 1502 |
| FrontierCode | 47.5% | — |
| CursorBench | 41.7% | — |
| GSO | 76.5% | — |
| WeirdML | 89.4% | — |
| MirrorCode | 20% | — |
| ALE-Bench | 2,177 | — |
Agentic & Tool Use GPT-5.6 Sol leads
GPT-5.6 Sol: 50.3 (#7), Qwen3.8 Max: 45.4 (#14)
| Benchmark | GPT-5.6 Sol | Qwen3.8 Max |
|---|---|---|
| APEX-Agents | 51.4% | 63.3% |
| τ²-bench Banking | 46.9% | 55.1% |
| GDP.pdf | 30.7% | 23.2% |
| OSWorld 2.0 | 27.3% | — |
| PostTrainBench | 36.2% | — |
| BALROG | 60% | — |
| GBAEval | 52.6% | — |
| LMArena Search | 1257 | — |
| Vending-Bench 2 | 9,619 | — |
Reasoning GPT-5.6 Sol leads
GPT-5.6 Sol: 74.8 (#8), Qwen3.8 Max: 54.4 (#26)
| Benchmark | GPT-5.6 Sol | Qwen3.8 Max |
|---|---|---|
| NYT Connections (extended) | 93.8% | 88.3% |
| CritPt | 32.3% | 20% |
| Chess Puzzles | 64% | 40% |
| LMArena Hard Prompts | 1484 | 1496 |
| Mystery Game Puzzles | 58% | 38% |
| DTBench | 96% | 92% |
| LMCA | 59.2% | 46.2% |
| Epoch Capabilities Index | 161.66 | 156.41 |
| ARC-AGI-2 | 92.5% | — |
| SimpleBench | 71.7% | — |
| Kagi LLM Benchmark | 67% | — |
| ARC-AGI-1 | 97.5% | — |
| EnigmaEval | 37.1% | — |
| EBR-Bench | 44.8% | — |
| Surface Evolver Bench | 93.1% | — |
| Bench to the Future 3 | 0.14 | — |
Math GPT-5.6 Sol leads
GPT-5.6 Sol: 85.6 (#9), Qwen3.8 Max: 73.2 (#20)
| Benchmark | GPT-5.6 Sol | Qwen3.8 Max |
|---|---|---|
| FrontierMath (Tiers 1-3) | 89.1% | 74.7% |
| FrontierMath Tier 4 | 82.9% | 46.3% |
| OTIS Mock AIME 2024-2025 | 100% | 100% |
| ProofBench | 83% | 58% |
| LMArena Math | 1474 | 1499 |
| FrontierMath Erdős | 0% | — |
Knowledge GPT-5.6 Sol leads
GPT-5.6 Sol: 64.3 (#18), Qwen3.8 Max: 61.7 (#27)
| Benchmark | GPT-5.6 Sol | Qwen3.8 Max |
|---|---|---|
| GPQA Diamond | 93.5% | 92.7% |
| SimpleQA Verified | 69.7% | 47.3% |
| LMArena Expert | 1516 | 1507 |
| Vectara Hallucination Rate | 12.4% | — |
Multimodal GPT-5.6 Sol leads
GPT-5.6 Sol: 48.6 (#9), Qwen3.8 Max: 37.2 (#75)
| Benchmark | GPT-5.6 Sol | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | 1281 | 1314 |
| Furniture Assembly | 56.7% | 20% |
| Blueprint-Bench 2 | 33.6% | — |
| LMArena Document | 1483 | — |
Multilingual Qwen3.8 Max leads
GPT-5.6 Sol: 55.3 (#32), Qwen3.8 Max: 56.7 (#18)
| Benchmark | GPT-5.6 Sol | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1452 | 1472 |
| LMArena Chinese | 1527 | 1538 |
| LMArena French | 1477 | 1503 |
| LMArena German | 1476 | 1483 |
| LMArena Japanese | 1471 | 1467 |
| LMArena Korean | 1442 | 1461 |
| LMArena Russian | 1468 | 1481 |
| LMArena Spanish | 1441 | 1492 |
Instruction Following Too close to call
GPT-5.6 Sol: 77.7 (#16), Qwen3.8 Max: 77.6 (#17)
| Benchmark | GPT-5.6 Sol | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1482 | 1479 |
Long Context Too close to call
GPT-5.6 Sol: 45.4 (#42), Qwen3.8 Max: 45.6 (#31)
| Benchmark | GPT-5.6 Sol | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1480 | 1489 |
Writing & Preference GPT-5.6 Sol leads
GPT-5.6 Sol: 73.3 (#12), Qwen3.8 Max: 67.1 (#30)
| Benchmark | GPT-5.6 Sol | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1457 | 1483 |
| LMArena Creative Writing | 1448 | 1479 |
| LMArena Multi-Turn | 1460 | 1489 |
| EQ-Bench Creative Writing | 1972 | — |
| EQ-Bench 4 | 1250 | — |
Frequently asked questions
Is GPT-5.6 Sol better than Qwen3.8 Max?
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 56.8 on the Noometry Index. Qwen3.8 Max costs 2.7× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Which is cheaper, GPT-5.6 Sol or Qwen3.8 Max?
Qwen3.8 Max is cheaper. It lists at $2 per million input tokens and $6 per million output tokens; GPT-5.6 Sol lists at $4 and $20.
Is GPT-5.6 Sol or Qwen3.8 Max better for coding?
GPT-5.6 Sol scores higher on coding benchmarks: 65.1 versus 53.5 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Sol does, with 1.05M tokens against 1M.
How many benchmarks do GPT-5.6 Sol and Qwen3.8 Max share?
39 benchmarks have published results for both models. GPT-5.6 Sol has 65 scored results on Noometry and Qwen3.8 Max has 39.