On this page
Alibaba's official Qwen account @Alibaba_Qwen announced at 2026-11-11 02:01:09 UTC (2026-11-11 10:01:09 Asia/Shanghai) that Qwen4-Max is now in preview on Alibaba's Token Plan. Official documentation lists the formal model ID as qwen4-max-preview. The announcement is a preview opening, not a generally available or permanent release, and official materials for this preview do not present Qwen4 27B or open weights.
*Official marketing poster from the @Alibaba_Qwen announcement post, marked "Now in Preview." This is a marketing poster, not a benchmark or evaluation chart.*
The official post says:
> Qwen4 is coming. > > Deeper knowledge. Richer world understanding. Stronger coding and agentic capabilities. > > Built for real-world tasks — from complex coding and knowledge work to front-end builds and professional software workflows. > > Qwen4-Max is now in preview on Alibaba’s Token Plan. You can also get early access to the new model on QwenWork, Qoder, Qwen APP, and QwenStudio.
The developer account @QwenDevs later quoted that official post at 2026-11-11 02:28:40 UTC (2026-11-11 10:28:40 Asia/Shanghai) with the line "Qnow! Show us what you are building!" That quote is not a second official announcement; the announcement clock remains the @Alibaba_Qwen post.
Preview window
The preview window comes from official documentation, not from the announcement post's clock. QwenCloud documentation, and in parallel the international Alibaba Cloud English help overview (Last Updated: Nov 11, 2026), state that the preview period runs from 10:00 on November 11, 2026 to 24:00 on November 25, 2026 (UTC+8). Official wording says no changes will be made during this period, and the model will be retired after the preview ends. That freeze applies to this preview model. China-site official documentation records the same window and the same retirement after the preview ends.
Token Plan access
Official documentation states that Qwen4-Max-Preview is available through Token Plan only. Pay-as-you-go access is not supported. The QwenCloud Latest model page says access is through Token Plan Personal or Token Plan Team, by following the corresponding guide to obtain a dedicated API key, using the dedicated Base URL https://token-plan.maas.qwencloudapi.com, and setting the model ID to qwen4-max-preview.
Credits under preview benefits
Official documentation places a limited-time Credits promotion under preview benefits. During that promotion, Credits consumption is as low as 20% of the standard rate (5x usage). From 22:00 to 08:00 UTC+8, an additional 80% off stacks on that rate, which official English wording states is 4% of the standard rate. Official Chinese wording is 2 折, equivalent to 5 times the usage, and 2 折再享 2 折, or 原标准的 0.4 折. The denominator is Credits consumption relative to the standard rate, not a token unit price, not the plan's monthly fee, and not a free allowance. Official pages do not publish an absolute Credits-per-million-tokens figure for this model.
Official documentation also states that the Credits consumed per request are dynamically determined by model type, token usage, thinking mode, and tool calls. That mechanism is not a rule that 1 Credit equals 1 Token.
The offer sits under preview benefits. Official text does not give a separate discount end date. Documentation says promotion details and validity periods follow the latest page content; the international Alibaba Cloud English overview also points to Alibaba Cloud notifications. The clock 24:00 on November 25, 2026 (UTC+8) is the official end of the preview window, when the model is retired.
This Credits offer is not the "Limited-time night discounts" / 限时夜间折扣 list that names qwen3.8-max, qwen3.8-flash, and specified DeepSeek IDs. That night list does not include qwen4-max-preview. It is also not the new-user limited plan price (through 2026-11-30 23:59:59 UTC+8, for Lite and Essential), not Credit Pack pricing, and not the November 9 plan-credit and Pro price change.
Alias and consoles
China-site documentation and the international Alibaba Cloud English help overview both list a billing alias. Official English wording on the international overview is: "The alias of qwen4-max-preview is qwen4.0-max-preview. Both IDs can be used; bills uniformly show qwen4.0-max-preview." Official Chinese wording on the China-site overview is that the alias of qwen4-max-preview is qwen4.0-max-preview, both IDs can be called, and bills uniformly display qwen4.0-max-preview. The international QwenCloud Token Plan Personal overview does not list this alias.
These consoles are not the same surface. The international Alibaba Cloud English overview states: "Token Plan Personal Edition is currently available only in the Singapore region." China-site documentation states that Token Plan Personal currently supports only 华北2（北京）. International QwenCloud, international Alibaba Cloud English help, and the China-site 华北2（北京） console remain separate. English pages under help.aliyun.com/en are China-site English translations, with China (Beijing) and RMB prices, not the international dollar documentation on www.alibabacloud.com.
The primary source for the international preview window and Credits wording remains QwenCloud documentation, and the same window and 20% / 4% figures also appear on the international Alibaba Cloud English help overview.
How official docs list the model
The international Alibaba Cloud English deep-thinking page (Last Updated: Nov 11, 2026) states: "Qwen4-Max series (preview, Hybrid thinking mode, thinking mode enabled by default): qwen4-max-preview (Token Plan only)."
The China-site deep-thinking page states: 「千问4-Max 系列（预览版，混合思考模式，默认开启思考模式）：qwen4-max-preview（仅 Token Plan）」.
The China-site model list carries two qwen4-max-preview cards, worded 「复杂推理与编程（Token Plan）」 and 「视觉理解（Token Plan）」.
Specs on the Latest model page
The QwenCloud Latest model page lists, as model specifications rather than benchmark scores, a 1M-token context length and a maximum output of 393,216 tokens, shared by thinking and the final answer. Thinking is enabled by default. Starting with this ID, max_tokens is equivalent to max_completion_tokens, and both limit thinking plus the final answer combined. The page also lists image and video understanding; Function Calling, web search, and built-in tools; JSON Object and JSON Schema structured outputs; and compatibility with OpenAI Chat Completions, Responses, and Anthropic Messages APIs.
The same page describes Qwen4-Max-Preview as the next-generation flagship multimodal model in the Qwen family, combining text generation, image and video understanding, and deep reasoning, and designed for coding, office productivity, frontend design, and complex agentic tasks. Those lines, like the official post's "deeper knowledge" and "stronger coding and agentic capabilities," are official self-description, not published benchmark scores. Official materials for this preview do not include a benchmark chart or evaluation table.
Other product names in the official post
The official post says the new model can be used for early access on QwenWork, Qoder, Qwen APP, and QwenStudio. That statement is attributed to the official post.
The Token Plan pricing page linked from the official post is the announcement's outbound Token Plan entry. It is not the source for the model ID, the preview window, or the Credits offer. The static Supported Models table on that page still showed "Loading..." and the page source did not list qwen4-max-preview.
The QwenCloud models changelog does not carry a standalone qwen4-max-preview release entry in the visible body. The Token Plan credits changelog dated November 9 is the plan-credit and Pro price change, a separate topic. Official blog surfaces at qwen.ai/blog and qwenlm.github.io are not the publication pages for a standalone article on this preview.
Related Token Plan coverage
Qwen4-Max's time-limited preview, with an explicit start and end and retirement after the window, plus the Credits consumption offer under preview benefits, is distinct from Alibaba Cloud Model Studio Token Plan Softens Night Credits for Qwen and DeepSeek, which covers the night Credits promotion for specified Qwen3.8 and DeepSeek models and does not include qwen4, and from Alibaba Cloud Raises Token Plan Credits; Pro Is $43 Internationally and ¥312 in China, which covers the November 9 plan-credit and Pro price change.