What is Qwen4-Max?

Qwen4-Max is the Qwen Team's flagship Qwen4 model, announced on 16 November 2026. Official pages describe a 2.4-trillion-parameter mixture-of-experts backbone, 71 billion parameters activated per token, native vision, and a 1 million token context window. It is available on QwenCloud via API or a Token Plan, with official pay-as-you-go rates on the live model page.

What you can use Qwen4-Max for

Qwen4-Max is the Qwen Team's flagship model in the new Qwen4 series, announced on 16 November 2026. The official blog describes it as an AI teammate for work, creativity, and discovery, built on the new Qwen4 architecture, with a 2.4-trillion-parameter backbone, native vision capabilities, and a one-million-token context window. It is a sparse mixture-of-experts model, with 71 billion parameters activated per token and an additional 328 billion N-gram Embedding parameters.

The official QwenCloud model page describes it as a multimodal model for text generation, image and video understanding, and deep reasoning, aimed at coding, office productivity, frontend design, and complex agentic tasks. Official input types are text, image, and video; official output is text. Listed features include function calling, context cache, structured outputs, batches, web search, prefix completion, and fine-tuning. The same official page says the model is compatible with the OpenAI and Anthropic APIs. This profile is based on official pages, not on a signed-in or billed API test.

A first workflow to try

Start from the official QwenCloud model page for qwen4-max, or from the access routes named in the blog: the QwenCloud API, the QwenCloud Token Plan, or the QwenWork, Qoder, and QwenStudio products. The official Qwen blog is dated 16 November 2026 and is signed by the Qwen Team. A reasonable first task is a bounded coding or document job that stays well inside the official 1 million token context, then compare the billed tokens and cache behavior with the live official rates.

The blog says smaller models in the Qwen4 family are coming to open source. Those weights were not published with this announcement, so this profile does not treat Qwen4-Max as an open-weight model.

Plans and access

On 16 November 2026, the official QwenCloud model page for qwen4-max, in the Singapore region, listed these pay-as-you-go rates per 1 million tokens: input $1.5, output $4.5, implicit-cache input $0.075, 5-minute explicit cache creation $1.875, 1-hour explicit cache creation $3, and explicit cache read $0.075. The same page lists a 1 million token context window, a maximum input of 991K tokens (983K in thinking mode), a maximum output of 393K tokens, a maximum reasoning budget of 262K tokens, and a 5,000,000 tokens-per-minute rate limit. A separate official QwenCloud API pricing table failed to load in this session, so the figures above come from the model page. Confirm them there before you budget.

QwenCloud also sells a Token Plan subscription. The official Token Plan page showed a new-customer offer from $3 and stated that Individual Lite and Essential plans are normally from $6 per month, with an extra 50% off during an event. Some plan-card totals and the supported-model list did not load in this session, so those missing card prices are not invented here. The official blog also lists QwenWork, Qoder, and QwenStudio as places to use the model, and the Qwen site is marked as powered by Alibaba Cloud. The QwenCloud pages checked for this profile carry a 2026 copyright line for Intelligent Cloud Computing (Singapore) Private Limited. On 16 November 2026, Qwen4-Max was listed on OpenRouter; that listing is not an official Qwen price list.

What stands out

  • Official 2.4-trillion-parameter Qwen4 backbone, with 71 billion parameters activated per token and native vision, as stated on the official blog
  • Official 1 million token context window, with documented max input, output, and reasoning limits on the QwenCloud model page
  • Official multimodal inputs (text, image, and video) and text output, plus function calling, cache, structured outputs, batches, and web search
  • Official access on QwenCloud through a Token Plan or API, and also on QwenWork, Qoder, and QwenStudio, per the official blog
  • Compatibility with the OpenAI and Anthropic APIs, as stated on the official QwenCloud model page

Before you choose

  • Official API rates above come from the QwenCloud model page; a separate official API pricing table failed to load in this session
  • Some Token Plan card totals did not load on the official subscription page in this session, so those missing prices are not listed here
  • Smaller Qwen4 models are described as coming to open source and were not released with this announcement; this is not a hands-on API test

Sources & further reading

Based on the official documentation below. Features and access can change; check the provider before making a decision.

  1. Qwen official blog: Qwen4-Max: Your New Teammate Has Arrived (16 November 2026, Qwen Team) (opens in a new tab)
  2. Official QwenCloud model page for qwen4-max, checked 16 November 2026 (Singapore region) (opens in a new tab)
  3. Official QwenCloud Token Plan page, checked 16 November 2026 (opens in a new tab)
  4. Official QwenCloud API pricing page, checked 16 November 2026 (model table failed to load in this session) (opens in a new tab)
  5. OpenRouter public listing for qwen/qwen4-max, checked 16 November 2026 (opens in a new tab)
Something out of date or incorrect?Report a correction

Was this page useful?

Help us make our coverage more useful. This is feedback on the page, not a rating of the tool.

Share this page