
Qwen4-Max: what Alibaba launched and who should try it
5 min read
Last updated November 17, 2026. How we cover news: we write from the official announcement and docs (linked below) and point you to the tools on Unloc. Not sponsored.

Short answer: Alibaba announced Qwen4-Max on November 16, 2026. It is the new flagship Qwen model: a 2.4-trillion-parameter mixture-of-experts model with native image and video understanding and a one-million-token context window, built for coding, office work and long-running agent tasks. It is available now through Alibaba's QwenCloud API and Token Plan subscription and in Qwen's own apps, but unlike some earlier Qwen releases it is not open weights.
Sponsored
Your tool here: sponsor this list
Fixed $149/month on Draft with AI. A subscription, not Claim #1.
Pin this category · $149/moIf you build with AI models, pay for API usage, or want a cheaper model to plug into a coding agent such as Claude Code or Cursor, this launch is worth a look. Below we cover what Alibaba announced, what is new in the model, where you can use it, what it costs at launch, how it compares with ChatGPT, Claude and Gemini, and what it means when you are choosing tools. We worked from the Qwen team's launch post, Alibaba Cloud's announcement and the QwenCloud developer docs.
What Alibaba announced
In the Qwen team's launch post dated November 16, 2026, Alibaba describes Qwen4-Max as "your new AI teammate for work, creativity, and discovery". Alibaba Cloud's announcement the same day calls it the most capable model in the Qwen family so far. According to the QwenCloud model changelog, a preview version had been available to Token Plan subscribers since November 11; the full model went on general release on November 16.
Alibaba also says smaller open-weight models from the Qwen4 series will follow for developers, but it has not given a date. Qwen4-Max itself is a hosted model you use through Alibaba's services.
What's new in Qwen4-Max
- A bigger but more efficient model. Alibaba says Qwen4-Max has 2.4 trillion parameters but activates only 71 billion per token, 25% fewer than Qwen3.8-Max at the same scale, using a new Qwen4 architecture.
- Native vision and a long context. The QwenCloud model docs list text, image, video and PDF understanding with text output, a one-million-token context window, and up to 262,144 tokens for thinking plus the final answer.
- Thinking on by default. Reasoning is enabled by default, with low, high and xhigh effort settings, plus tool calling, structured outputs, web search, a code interpreter and MCP support.
- Live Delivery for long tasks. Alibaba says agents running on Qwen4-Max can hand over a usable first version early and keep improving it, so you can review progress and change direction mid-task. It has open-sourced a Live Delivery plugin on GitHub.
- Lower cost per task. On Work360, Alibaba's own benchmark of professional tasks, the Qwen team says average cost per task fell from $4.43 with Qwen3.8-Max to $1.67, a 62% cut, helped by using fewer tokens.
Alibaba also reports strong scores on coding, agent and visual reasoning benchmarks and says the model ranked 6th on the Code Arena leaderboard at launch. These are the company's own launch figures; independent testing will show how it holds up on your own tasks.
Where you can use it
Alibaba lists four ways in:
- QwenCloud API (model ID
qwen4-max), which supports OpenAI-style Chat Completions and Responses requests and Anthropic Messages requests, so many existing apps can switch with a configuration change. - Token Plan, a QwenCloud subscription billed in credits. According to the Token Plan docs, one plan works across coding and agent tools including Claude Code, Cursor, Codex, Qwen Code, Qoder and OpenClaw. A Team edition adds seat management and says your data will not be used for model training.
- Qwen Studio, Qwen's chat app on the web.
- QwenWork and Qoder: QwenWork is Qwen's AI work app for reports, decks and other files, and Qoder is an agentic coding platform that Alibaba says it co-designed the model with.
What it costs
Pricing is part of the pitch. Alibaba says it lowered API prices compared with Qwen3.8-Max, with the biggest cut on cached input. At launch, the QwenCloud pricing page listed Qwen4-Max in its Singapore (International) region at $1.50 per million input tokens and $4.50 per million output tokens, with cached input reads at $0.075 per million, as checked on November 17, 2026. Prices differ by region and can change, so check the live pricing page before you budget. Token Plan tiers are priced separately in credits.
How it compares with ChatGPT, Claude and Gemini
For most people choosing a chat assistant, ChatGPT, Claude and Gemini remain the easier picks: they have polished apps, wide integrations and large communities, and Qwen4-Max does not change that on its own. Qwen4-Max is more interesting if you buy model access directly. Its one-million-token context, compatibility with OpenAI and Anthropic style APIs and the per-task savings Alibaba reports make it a candidate for long-document work, agents and coding at lower cost.
The trade-offs are practical. Qwen4-Max is closed, so you cannot run it on your own hardware the way you can with open-weight Qwen models. You also need to be comfortable with Alibaba Cloud as a provider: QwenCloud serves the model from regions including Singapore, Beijing, Frankfurt, Tokyo and US Virginia, so check which region and data terms fit your organisation. If you mainly want a coding agent, our Claude Code vs Cursor comparison covers the tools you would plug it into.
What it means if you're choosing tools
Qwen4-Max adds another credible option at the top end, and the pressure on price is the headline for buyers. If you spend heavily on API calls for agents or long documents, it is worth running your own tests against your current model, either through the pay-as-you-go API (QwenCloud's docs say new users in its Singapore region get a free token quota) or through a Token Plan in a coding tool you already use. If you are an everyday user happy with your current assistant, there is no reason to switch today; see our Claude vs ChatGPT comparison or browse the ChatGPT, Claude and Gemini listings on Unloc.
FAQ
What is Qwen4-Max?
Qwen4-Max is Alibaba's flagship Qwen AI model, announced on November 16, 2026. It is a 2.4-trillion-parameter mixture-of-experts model that understands text, images and video, has a one-million-token context window and is designed for coding, professional work and long-running agent tasks.
Is Qwen4-Max open source?
No. Qwen4-Max is a hosted model available through QwenCloud and Alibaba's apps. Alibaba says smaller open-weight models from the Qwen4 series will be released later, but it has not given a date.
Where can I use Qwen4-Max?
Alibaba says it is available through the QwenCloud API and Token Plan, and in Qwen Studio, QwenWork and Qoder. Token Plan also lets you use it inside tools such as Claude Code, Cursor and Codex.
How much does Qwen4-Max cost?
At launch, QwenCloud listed $1.50 per million input tokens and $4.50 per million output tokens in its Singapore (International) region, with lower rates for cached input. Prices vary by region and can change, so check QwenCloud's live pricing page.
Is Qwen4-Max better than ChatGPT or Claude?
Alibaba's own benchmarks show strong coding and agent results at a lower cost per task than its previous model, but those are vendor figures. For everyday chat, ChatGPT and Claude are still the more polished apps; for API work, test Qwen4-Max on your own tasks before switching.
Submit a tool if it belongs on the board. Listings are reviewed before they go live.
Share your learnings on socials
Other resources
View all
DeepSeek vs ChatGPT: which AI assistant should you use?
Pick ChatGPT for the more complete assistant app with images, voice, projects and agents. Pick DeepSeek for chat with search and deep thinking, a low-cost OpenAI-compatible API and MIT-licensed open weights, and read its privacy policy first.

ChatGPT Wallet: how agent payments work and who can use it
OpenAI announced ChatGPT Wallet on November 13, 2026: agents in ChatGPT can pay for you with a one-purchase payment token after you approve with Face ID or a fingerprint. Here's how it works, who can use it, what shops need to change and how it compares with Gemini and Claude.

Best Notion alternatives: free and paid workspace apps
The best Notion alternative depends on the job: ClickUp and Superhuman Docs suit all-in-one team workspaces, Confluence suits company wikis, Obsidian and Craft suit personal notes, AppFlowy and Anytype suit open source and private setups, and Linear and Todoist suit tasks and projects.