Is Qwen4-Max-Preview state of the art?Pending benchmarks.Announced Nov 11, 2026; no independent results yet.
Qwen's preview multimodal model in the Qwen4 line: text generation, image and video understanding, and deep reasoning, with a 1M-token context and a 393,216-token maximum output shared by thinking and the final answer. Announced November 11, 2026. Available through Alibaba Cloud Token Plan only (model qwen4-max-preview, alias qwen4.0-max-preview) during a preview window from 10:00 November 11 to 24:00 November 25, 2026 (UTC+8); pay-as-you-go is not supported, and Alibaba says the model will be retired after the preview ends.
The details so far
- Lab
- Alibaba / Qwen
- Announced
- Nov 11, 2026
- API
- Live · qwen4-max-preview
- Open weights
- Closed · No public weight release stated
- Type
- Multimodal model for text generation, image and video understanding, and deep reasoning. Thinking is on by default; reasoning_effort can be set to low, medium, or xhigh (Alibaba Cloud Model Studio docs)
- Context
- 1M-token context; maximum output 393,216 tokens, shared by thinking and the final answer (Alibaba Cloud Model Studio docs)
- Tools
- Function calling, web search, built-in tools, and structured outputs (JSON Object and JSON Schema). Compatible with OpenAI Chat Completions, Responses, and Anthropic Messages APIs (Alibaba Cloud Model Studio docs)
- Hosted access
- Alibaba Cloud Token Plan only (Personal or Team), model qwen4-max-preview (alias qwen4.0-max-preview). Dedicated Token Plan base URL https://token-plan.ap-southeast-1.maas.aliyuncs.com. Preview window: 10:00 November 11, 2026 to 24:00 November 25, 2026 (UTC+8); no changes during the window, then retired. Pay-as-you-go is not supported
- Pricing
- Billed in Token Plan Credits, not a published per-token list price. Token Plan Personal (Singapore region): Lite $6/month (limited-time new-user $3/month through November 30, 2026 UTC+8), Essential $10 ($5 new-user), Standard $18, Pro $43, with monthly quotas of 20,000 to 200,000 Credits. During the preview promotion, Credits consumption is as low as 20% of the standard rate, and 8% of the standard rate for calls between 22:00 and 08:00 UTC+8 (Token Plan Personal docs)
- Plan terms
- Token Plan Personal is documented as a self-serve subscription. Official docs limit it to interactive use in coding and agent tools, not automation scripts or custom application backends (Token Plan Personal docs)
- License
- Not stated
- Related
- Separate from Qwen Max, Qwen3 Max, Qwen3 Max Preview, Qwen3.5 Max Preview, Qwen3.6 Max Preview, Qwen3.7 Max, Qwen3.8 Max, and Qwen VL Max pages
Tags marked unsure are not stated on the lab's own pages.
What would let us call it
Verdicts use only independent boards we already track. Two consecutive snapshots are required before any crown.
- Publicly announced with a source
- Usable by the public (API or weights)
- Listed on an independent board we track
- Enough results for a rank in a category
- Holds across two snapshots (crowns only)
Sources
LABAlibaba Cloud Model Studio: qwen4-max-previewalibabacloud.com/help/en/model-studio/qwen4-max-preview · Nov 11, 2026LABAlibaba Cloud Model Studio: Token Plan (Personal Edition)alibabacloud.com/help/en/model-studio/token-plan-personal-overview · Nov 11, 2026LABAlibaba Cloud Model Studio: newly released models (2026-11-11)help.aliyun.com/en/model-studio/newly-released-models · Nov 11, 2026LABQwen on X: Qwen4-Max-Previewx.com/Alibaba_Qwen/status/2120335993643684315 · Nov 11, 2026Tracking continues until an independent board lists it.