Model comparison
Qwen3.8 Max vs Qwen3 Max
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 43.7 on the Noometry Index.
Last verified . 27 shared benchmarks.
Alibaba (Qwen)Summary
- They share 27 benchmarks with published results for both. Qwen3.8 Max scores higher in 8 categories and Qwen3 Max in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 38.7.
- The biggest single-benchmark swing is NYT Connections (extended): 88.3% for Qwen3.8 Max and 30.1% for Qwen3 Max.
- Qwen3 Max is cheaper at $1.20 / $6 per million input/output tokens, against $2 / $6 for Qwen3.8 Max.
- Qwen3.8 Max accepts more context: 1M tokens versus 262K.
Side by side
| Qwen3.8 Max | Qwen3 Max | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 56.8 | 43.7 |
| Released | 2026-08-02 | 2025-09-23 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 262K |
| Max output | 131K | 66K |
| Input $ / M tokens | $2 | $1.20 |
| Output $ / M tokens | $6 | $6 |
| Results tracked | 39 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.8 Max leads
Qwen3.8 Max: 53.5 (#29), Qwen3 Max: 43.0 (#93)
| Benchmark | Qwen3.8 Max | Qwen3 Max |
|---|---|---|
| LMArena Coding | 1502 | 1456 |
| DeepSWE | 57.5% | — |
| LMArena WebDev | 1674 | — |
| FrontierSWE | 17.8% | — |
| SciCode | 53.2% | — |
| ALE-Bench | — | 370.45 |
Agentic & Tool Use Not comparable
Qwen3.8 Max: 45.4 (#14), Qwen3 Max: —
| Benchmark | Qwen3.8 Max | Qwen3 Max |
|---|---|---|
| APEX-Agents | 63.3% | — |
| τ²-bench Banking | 55.1% | — |
| GDP.pdf | 23.2% | — |
| Vending-Bench 2 | — | 71.56 |
Reasoning Qwen3.8 Max leads
Qwen3.8 Max: 54.4 (#26), Qwen3 Max: 22.6 (#190)
| Benchmark | Qwen3.8 Max | Qwen3 Max |
|---|---|---|
| NYT Connections (extended) | 88.3% | 30.1% |
| Chess Puzzles | 40% | 4% |
| LMArena Hard Prompts | 1496 | 1448 |
| Mystery Game Puzzles | 38% | 5% |
| DTBench | 92% | 82.1% |
| LMCA | 46.2% | 28.3% |
| Epoch Capabilities Index | 156.41 | 142.38 |
| Kagi LLM Benchmark | — | 72.5% |
| CritPt | 20% | — |
Math Qwen3.8 Max leads
Qwen3.8 Max: 73.2 (#20), Qwen3 Max: 38.7 (#131)
| Benchmark | Qwen3.8 Max | Qwen3 Max |
|---|---|---|
| FrontierMath (Tiers 1-3) | 74.7% | 18.9% |
| OTIS Mock AIME 2024-2025 | 100% | 73.3% |
| LMArena Math | 1499 | 1446 |
| FrontierMath Tier 4 | 46.3% | — |
| ProofBench | 58% | — |
| MATH Level 5 | — | 97.1% |
Knowledge Qwen3.8 Max leads
Qwen3.8 Max: 61.7 (#27), Qwen3 Max: 48.1 (#78)
| Benchmark | Qwen3.8 Max | Qwen3 Max |
|---|---|---|
| GPQA Diamond | 92.7% | 72.6% |
| SimpleQA Verified | 47.3% | 48.7% |
| LMArena Expert | 1507 | 1455 |
Multimodal Not comparable
Qwen3.8 Max: 37.2 (#75), Qwen3 Max: —
| Benchmark | Qwen3.8 Max | Qwen3 Max |
|---|---|---|
| LMArena Vision | 1314 | — |
| Furniture Assembly | 20% | — |
Multilingual Qwen3.8 Max leads
Qwen3.8 Max: 56.7 (#18), Qwen3 Max: 53.7 (#62)
| Benchmark | Qwen3.8 Max | Qwen3 Max |
|---|---|---|
| LMArena Non-English | 1472 | 1429 |
| LMArena Chinese | 1538 | 1478 |
| LMArena French | 1503 | 1449 |
| LMArena German | 1483 | 1463 |
| LMArena Japanese | 1467 | 1397 |
| LMArena Korean | 1461 | 1399 |
| LMArena Russian | 1481 | 1428 |
| LMArena Spanish | 1492 | 1462 |
Instruction Following Qwen3.8 Max leads
Qwen3.8 Max: 77.6 (#17), Qwen3 Max: 74.8 (#87)
| Benchmark | Qwen3.8 Max | Qwen3 Max |
|---|---|---|
| LMArena Instruction Following | 1479 | 1419 |
Long Context Qwen3.8 Max leads
Qwen3.8 Max: 45.6 (#31), Qwen3 Max: 41.6 (#134)
| Benchmark | Qwen3.8 Max | Qwen3 Max |
|---|---|---|
| LMArena Longer Query | 1489 | 1438 |
| Fiction.LiveBench | — | 66.7% |
| CL-bench | — | 14.5% |
Writing & Preference Qwen3.8 Max leads
Qwen3.8 Max: 67.1 (#30), Qwen3 Max: 62.4 (#76)
| Benchmark | Qwen3.8 Max | Qwen3 Max |
|---|---|---|
| LMArena Text | 1483 | 1439 |
| LMArena Creative Writing | 1479 | 1402 |
| LMArena Multi-Turn | 1489 | 1446 |
Frequently asked questions
Is Qwen3.8 Max better than Qwen3 Max?
Qwen3.8 Max is the stronger model overall, scoring 56.8 to 43.7 on the Noometry Index.
Which is cheaper, Qwen3.8 Max or Qwen3 Max?
Qwen3 Max is cheaper. It lists at $1.20 per million input tokens and $6 per million output tokens; Qwen3.8 Max lists at $2 and $6.
Is Qwen3.8 Max or Qwen3 Max better for coding?
Qwen3.8 Max scores higher on coding benchmarks: 53.5 versus 43.0 in the Noometry coding category.
Which has the bigger context window?
Qwen3.8 Max does, with 1M tokens against 262K.
How many benchmarks do Qwen3.8 Max and Qwen3 Max share?
27 benchmarks have published results for both models. Qwen3.8 Max has 39 scored results on Noometry and Qwen3 Max has 33.