Skip to content

ModelsAlibaba QwenUpdated

Qwen4-Max API launches: Alibaba Qwen's Qwen 4 flagship is generally available at $1.50/$4.50 per million tokens

Qwen4-Max at a glance

The Qwen4-Max API is live on QwenCloud: Alibaba Qwen made its Qwen 4 flagship generally available on November 16, 2026. The Qwen4-Max price is $1.50/$4.50 per million tokens, with native vision, a 1M-token context and a score of 43 on the Artificial Analysis Intelligence Index as of November 16, 2026.

Current standing

As of

Rank
#15of 47 models

Qwen4 Max (XHigh) · 42.9 on the Artificial Analysis Intelligence Index v5.0.1 (opens in a new tab)

Recorded at publication
43 (v5.0.1)

as of

Current leader
Claude Fable 5.2 (Anthropic) · 64.6

Computed from our daily snapshot; it changes as new models are measured. Qwen4-Max API launches: Alibaba Qwen's Qwen 4 flagship is generally available at $1.50/$4.50 per million tokens

Ask your AI about this story

Use this story in your business: Qwen4-Max API launches: Alibaba Qwen's Qwen 4 flagship is generally available at $1.50/$4.50 per million tokens (Alibaba Qwen). Ask your AI to explain what happened, how it could affect your business and how to put it to work. It also weighs price, speed, intelligence and the lab behind it. The link to the story is included.

What changed with Qwen4-Max

Alibaba Qwen made Qwen4-Max generally available on November 16, 2026, five days after its Token Plan preview opened. The Qwen4-Max API is live on QwenCloud with pay-as-you-go billing, and the model also runs in Qwen Studio, QwenWork and Qoder. It is the first model on the new Qwen 4 architecture: a sparse mixture-of-experts design with a 2.4-trillion-parameter backbone, 71 billion parameters active per token and a further 328 billion in N-gram Embedding parameters. It takes text, image and video input, returns text, and works across a 1-million-token context window, with OpenAI- and Anthropic-compatible APIs and reasoning effort levels of low, high (the default) and xhigh.

The Qwen4-Max price undercuts the model it replaces: $1.50 per million input tokens and $4.50 per million output tokens on QwenCloud's Singapore international list (Beijing and Global deployments list $1.413 and $4.24), with cached input at $0.075, against $2 and $6 for Qwen3.8-Max. In Alibaba's own Work360 evaluation, the average cost per task fell from $4.43 to $1.67, a 62% drop, because the model uses roughly 24% to 49% fewer tokens at the same effort levels. One point needs stating plainly: Qwen4-Max is a closed model. There are no weights to download and no self-hosting; Alibaba says the smaller Qwen4 models will be released as open source later, with no date given.

Artificial Analysis scored Qwen4-Max at 43 on its Intelligence Index (v5.0.1) as of November 16, 2026, at its xhigh effort level, up from 34 for Qwen3.8-Max, with a cost of $3.27 per Index task against $10.02; as of November 16, 2026, Artificial Analysis had not published output-speed or latency figures for it. In Arena's leaderboard data of November 15, 2026, it debuted at #6 on the WebDev board with 1750 points (preliminary) and #14 on the Agent board. In Alibaba's own comparison chart, it scored 75.7 on DeepSWE v1.1, ahead of Claude Opus 5.5 at 74.2, and 52.69 on Work360 against 47.74 for Qwen3.8-Max, while trailing on Terminal-Bench 4.0 with 48.5 against Kimi K3.1's 67.2.

Alibaba Qwen's own Qwen4-Max benchmark chart: 16 bar charts comparing Qwen4-Max with Qwen3.8-Max, Claude Opus 5.5, GPT-6.1 Sol, GPT-6 Astra and Kimi K3.1 across coding, work and long-horizon tests, including DeepSWE v1.1 at 75.7 and Terminal-Bench 4.0 at 48.5.
Official imageImage: Alibaba Qwen (opens in a new tab)

Qwen4-Max benchmarks and pricing

Values are copied from the linked source on the date shown. They are not IA Compa measurements.

Artificial Analysis Intelligence Indexas of · Source for Artificial Analysis Intelligence Index: artificialanalysis.ai (opens in a new tab)
43 (v5.0.1)
Input price per 1M tokensas of · Source for Input price per 1M tokens: docs.qwencloud.com (opens in a new tab)
$1.50
Output price per 1M tokensas of · Source for Output price per 1M tokens: docs.qwencloud.com (opens in a new tab)
$4.50
Context windowas of · Source for Context window: qwencloud.com (opens in a new tab)
1M tokens
Cost per Artificial Analysis Intelligence Index taskas of · Source for Cost per Artificial Analysis Intelligence Index task: artificialanalysis.ai (opens in a new tab)
$3.27
Work360 score (Alibaba's own evaluation)as of · Source for Work360 score (Alibaba's own evaluation): qwen.ai (opens in a new tab)
52.69
DeepSWE v1.1 (Alibaba's own evaluation)as of · Source for DeepSWE v1.1 (Alibaba's own evaluation): qwen.ai (opens in a new tab)
75.7

How it compares

Current models we have covered: released in the last six months, newest version of each line only. Each value is the model's latest recorded value. (Models)

Intelligence indexhigher is better
Input price (USD per 1M tokens)lower is better
Output price (USD per 1M tokens)lower is better
Context window (tokens)higher is better

Where to compare Qwen4-Max

Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.

Why Qwen4-Max matters for businesses

Why it matters: your team can rent a flagship-class model for agentic and document work at $1.50/$4.50 per million tokens, and Alibaba's own testing says each finished task costs about 62% less than on the previous flagship. Because the endpoints speak the OpenAI and Anthropic API formats, you can trial Qwen4-Max inside tools you already run and route routine work to it while keeping premium models for the hardest jobs. The trade-off is control: the weights are closed, so there is no self-hosted or offline option, your data travels to Alibaba Cloud regions, and your costs follow QwenCloud's list prices until the promised smaller open models arrive.

Put it to work: AI services

Get AI News by email

Verified AI stories, each with its official source, on Monday and Thursday mornings. Add breaking news to hear about the biggest ones as soon as we publish them.

Every story links its official source

Free, in English or Spanish

Unsubscribe with one click, any time

What to send

Qwen4-Max updates since publication

Dated changes since publication. The original article text above is kept as written on its date.

  1. Update (Updated)

    Benchmark update (November 16, 2026): Artificial Analysis has now published measured performance for Qwen4-Max — 33.2 minutes per Intelligence Index task and 139.46 seconds to the first answer token via Alibaba's API, among the slowest in its price tier. Its Intelligence Index (43) and Arena ranks are unchanged.

    artificialanalysis.ai (opens in a new tab)

Sources for this story

  1. Official
  2. Official
  3. Official
  4. Benchmark
  5. Benchmark

Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.

  • #qwen4-max
  • #qwen4
  • #qwen
  • #alibaba-qwen
  • #alibaba
  • #mixture-of-experts
  • #ai-api
  • #multimodal
  • #long-context
  • #coding-agents

Want this working in your business?

We turn new AI capabilities into agents, voice receptionists, and automations you can actually use.