Model comparison
Qwen3.7 Max vs Qwen3.8 27B
Qwen3.7 Max is the stronger model overall, scoring 51.5 to 46.0 on the Noometry Index. Qwen3.8 27B costs 3.4× less per token, which makes it the better buy when Qwen3.7 Max's lead doesn't matter for your workload.
Last verified . 20 shared benchmarks.
Alibaba (Qwen)Summary
- They share 20 benchmarks with published results for both. Qwen3.7 Max scores higher in 6 categories and Qwen3.8 27B in 3 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.7 Max leads 62.4 to 37.1.
- The biggest single-benchmark swing is NYT Connections (extended): 85.1% for Qwen3.7 Max and 54.5% for Qwen3.8 27B.
- Qwen3.8 27B is cheaper at $0.99 / $1.49 per million input/output tokens, against $2.50 / $7.50 for Qwen3.7 Max.
- Qwen3.7 Max accepts more context: 1M tokens versus 262K.
- Qwen3.8 27B has downloadable open weights; the other is API-only.
Side by side
| Qwen3.7 Max | Qwen3.8 27B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 51.5 | 46.0 |
| Released | 2026-05-19 | 2026-08-14 |
| Weights | Proprietary | Open |
| Context window | 1M | 262K |
| Max output | 131K | 33K |
| Input $ / M tokens | $2.50 | $0.99 |
| Output $ / M tokens | $7.50 | $1.49 |
| Results tracked | 33 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Qwen3.7 Max: 50.4 (#45), Qwen3.8 27B: 50.5 (#44)
| Benchmark | Qwen3.7 Max | Qwen3.8 27B |
|---|---|---|
| LMArena WebDev | 1515 | 1593 |
| SciCode | 48.8% | 46.6% |
| LMArena Coding | 1498 | 1482 |
| SWE-bench Verified | 77.3% | — |
| ALE-Bench | 1,189 | — |
Agentic & Tool Use Qwen3.8 27B leads
Qwen3.7 Max: 22.1 (#135), Qwen3.8 27B: 32.9 (#57)
| Benchmark | Qwen3.7 Max | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | — | 47.5% |
| GBAEval | 0.4% | — |
Reasoning Qwen3.7 Max leads
Qwen3.7 Max: 49.2 (#38), Qwen3.8 27B: 41.0 (#54)
| Benchmark | Qwen3.7 Max | Qwen3.8 27B |
|---|---|---|
| NYT Connections (extended) | 85.1% | 54.5% |
| CritPt | 13.4% | 5.4% |
| LMArena Hard Prompts | 1483 | 1460 |
| DTBench | 92.3% | 88% |
| LMCA | 44% | 41.4% |
| Epoch Capabilities Index | 153.68 | 149.38 |
| ARC-AGI-2 | — | 42.4% |
| SimpleBench | 70.4% | — |
| ARC-AGI-1 | — | 87.5% |
| Chess Puzzles | 19% | — |
| EBR-Bench | 9.5% | — |
| Mystery Game Puzzles | 32% | — |
| Surface Evolver Bench | — | 45% |
Math Qwen3.7 Max leads
Qwen3.7 Max: 62.4 (#32), Qwen3.8 27B: 37.1 (#161)
| Benchmark | Qwen3.7 Max | Qwen3.8 27B |
|---|---|---|
| ProofBench | 26% | 16% |
| LMArena Math | 1490 | 1456 |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 34.1% | — |
| OTIS Mock AIME 2024-2025 | 95.6% | — |
Knowledge Qwen3.7 Max leads
Qwen3.7 Max: 61.6 (#28), Qwen3.8 27B: 41.6 (#109)
| Benchmark | Qwen3.7 Max | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1488 | 1482 |
| GPQA Diamond | 90.9% | — |
| SimpleQA Verified | 55.8% | — |
Multimodal Not comparable
Qwen3.7 Max: —, Qwen3.8 27B: 41.3 (#37)
| Benchmark | Qwen3.7 Max | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | — | 1271 |
Multilingual Qwen3.7 Max leads
Qwen3.7 Max: 56.9 (#15), Qwen3.8 27B: 53.7 (#60)
| Benchmark | Qwen3.7 Max | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1474 | 1430 |
| LMArena Chinese | 1530 | 1504 |
| LMArena Russian | 1484 | 1415 |
| LMArena French | — | 1465 |
| LMArena German | — | 1438 |
| LMArena Japanese | — | 1384 |
| LMArena Korean | — | 1393 |
| LMArena Spanish | — | 1448 |
Instruction Following Too close to call
Qwen3.7 Max: 76.7 (#38), Qwen3.8 27B: 75.8 (#53)
| Benchmark | Qwen3.7 Max | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1460 | 1439 |
Long Context Qwen3.7 Max leads
Qwen3.7 Max: 45.4 (#40), Qwen3.8 27B: 44.3 (#70)
| Benchmark | Qwen3.7 Max | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1482 | 1450 |
Writing & Preference Too close to call
Qwen3.7 Max: 65.0 (#54), Qwen3.8 27B: 65.8 (#43)
| Benchmark | Qwen3.7 Max | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1476 | 1441 |
| LMArena Creative Writing | 1449 | 1384 |
| LMArena Multi-Turn | 1481 | 1441 |
| EQ-Bench Creative Writing | — | 1671 |
| EQ-Bench 4 | 1110 | — |
Frequently asked questions
Is Qwen3.7 Max better than Qwen3.8 27B?
Qwen3.7 Max is the stronger model overall, scoring 51.5 to 46.0 on the Noometry Index. Qwen3.8 27B costs 3.4× less per token, which makes it the better buy when Qwen3.7 Max's lead doesn't matter for your workload.
Which is cheaper, Qwen3.7 Max or Qwen3.8 27B?
Qwen3.8 27B is cheaper. It lists at $0.99 per million input tokens and $1.49 per million output tokens; Qwen3.7 Max lists at $2.50 and $7.50.
Is Qwen3.7 Max or Qwen3.8 27B better for coding?
They score almost the same on coding (50.4 vs 50.5); test both on your own repository before choosing.
Which has the bigger context window?
Qwen3.7 Max does, with 1M tokens against 262K.
How many benchmarks do Qwen3.7 Max and Qwen3.8 27B share?
20 benchmarks have published results for both models. Qwen3.7 Max has 33 scored results on Noometry and Qwen3.8 27B has 31.