Gemini 4 Argon
Claude Opus 5.5
Muse Wonder
MiMo-V2.6-Pro
Kimi K3
GLM-5.3
GPT-5.5
ERNIE 5.1
DeepSeek-V4.1-Flash-20260910
Grok 4.20 Beta
amazon-nova-experimental-chat
Dola Seed 2.0 Pro
NVIDIA-Nemotron-3-Ultra-550B-A55B
Gemma 4 31B
Step 5 Preview
Qwen3-Max-Preview
Qwen3-Max-Preview is a proprietary model from Alibaba. It ranks #77 of 351 on the LMArena text leaderboard with a rating of 1439.
- Developer
- Alibaba Qwen
- Released
- Sep 8, 2025
- Context window
- —
- Price in / out, per 1M tokens
- —
- Weights
- Proprietary
- Inputs
- —
- Reasoning
- —
- Tool use
- —
The rating LMArena published at each update since Sep 8, 2025. Ratings move as votes come in and new models join.
Benchmarks
| Benchmark | Score bar | Accuracy | Rank | Cost / test |
|---|---|---|---|---|
| Vibe Code BenchCoding | 3.51% | 105 / 115 | $6.02 | |
| LiveCodeBenchCoding | 66.91% | 98 / 143 | $1.2 / $6 | |
| MedCodeMedicine | 31.37% | 94 / 107 | $1.2 / $6 | |
| MedQAMedicine | 87.37% | 56 / 92 | $1.2 / $6 | |
| GPQA DiamondReasoning | 77.78% | 77 / 138 | $1.2 / $6 | |
| MMLU-ProReasoning | 83.54% | 73 / 138 | $1.2 / $6 | |
| AIMEReasoning | 60.69% | 57 / 96 | $1.2 / $6 |
Vals AI, as of Oct 24, 2026.
LMArena, by category
| Leaderboard | Rating | Rank | Votes |
|---|---|---|---|
| TextLMArena | 1439 | 77 / 351 | 27.4K |
| CodingLMArena | 1456 | 90 / 346 | 5.4K |
| MathLMArena | 1446 | 77 / 339 | 1.5K |
| Creative writingLMArena | 1401 | 98 / 349 | 3.7K |
| Instruction followingLMArena | 1420 | 89 / 351 | 7.3K |
| Hard promptsLMArena | 1448 | 78 / 351 | 13.6K |
| Multi-turnLMArena | 1446 | 71 / 349 | 4.5K |
| Long queriesLMArena | 1438 | 81 / 330 | 6.2K |
| ExpertLMArena | 1455 | 76 / 305 | 1.3K |
| Legal and governmentIndustries | 1435 | 102 / 325 | 1.7K |
| Medicine and healthcareIndustries | 1468 | 43 / 325 | 1.5K |
| Business and financeIndustries | 1436 | 65 / 344 | 5.1K |
| ChineseLanguages | 1478 | 77 / 328 | 2.1K |
| SpanishLanguages | 1462 | 39 / 249 | 845 |
| GermanLanguages | 1462 | 37 / 261 | 596 |
| FrenchLanguages | 1450 | 87 / 252 | 438 |
| JapaneseLanguages | 1398 | 79 / 236 | 438 |
LMArena, CC BY 4.0, as of Oct 22, 2026. Where a model is rated at several effort levels, its best is shown (Qwen3-Max-Preview, Qwen3-Max (2025-09-23)).
Compare Qwen3-Max-Preview
Alibaba Qwen
Alibaba Cloud’s Qwen team releases a broad family of open-weight models, from small to frontier scale. Founded 2023 (Qwen), based in Hangzhou. qwenlm.github.io
Questions
Who makes Qwen3-Max-Preview?
Qwen3-Max-Preview is made by Alibaba Qwen, founded 2023 (Qwen) and based in Hangzhou.
How does Qwen3-Max-Preview rank?
#77 of 351 on LMArena’s text leaderboard, rated 1439 (95% interval 1434–1443) from 27,367 votes, as of Oct 22, 2026.
Is Qwen3-Max-Preview open source?
No. Qwen3-Max-Preview is proprietary and served through its provider’s API.