Alibaba
Qwen3.8 Omni Flash
Total Context
1M
Max Output
131.1K
Released
N/A
qwen4-maxQwen4-Max is Tongyi Qianwen's new-generation flagship multimodal model, combining text generation, image and video understanding, and deep reasoning. It is suited for coding, office work, front-end design, and complex agent tasks. The model supports web search, tool calling, and structured output, and is compatible with OpenAI and Anthropic APIs, helping developers build intelligent applications with multimodal understanding and task execution capabilities.
Context Window
1M tokens
Maximum Output
262.1K tokens
Release Date
Nov 16, 2026
Modalities
| Input Price | Output Price | Cache Read | Cache Create 5m |
|---|---|---|---|
| $1.425/M | $4.275/M | $0.0713/M | $1.781/M |
Reasoning
Tool calling
Temperature parameter
Attachments
Knowledge Base
Endpoint Protocols
Qwen4-Max combines native visual understanding with reasoning for projects that span multiple steps.
Interprets diagrams, screenshots, and video alongside written instructions, helping connect what is visible to the task at hand.
Can carry a plan through implementation, check intermediate results, and revise a deliverable in response to feedback.
A one-million-token context window lets the model follow evidence across long documents, code, and prior messages.
Teams can choose a lighter setting for straightforward work and reserve more reasoning for tasks that need deeper checks.
Use Qwen4-Max for coordinated code changes, document analysis, and design work grounded in visual references.
Trace existing components, implement a feature across the relevant files, and run targeted tests before review.
Reconcile figures and claims across PDFs, charts, and notes, then draft a brief with references and open questions.
Turn an approved design into a responsive interface and compare the rendered layout with the reference at several screen sizes.
Use reference images and a scene brief to build an editable 3D prototype in Blender, then inspect proportions and adjust the scene.
import OpenAI from "openai"
const client = new OpenAI({
apiKey: process.env.TOKENHUB_API_KEY,
baseURL: "https://us-api.tokenhub.com/v1",
})
const result = await client.chat.completions.create({})
console.log(result.choices[0]?.message?.content)Qwen4 Max
| Index score | ||
|---|---|---|
| Artificial Analysis Intelligence Index | Artificial Analysis broad capability aggregate | 42.9 |
Metrics sourced from Artificial Analysis
Context limits, reasoning effort, supported inputs, and API access for Qwen4-Max.
Qwen4-Max is Alibaba’s flagship multimodal Qwen4 model. It uses a mixture-of-experts architecture with 2.4 trillion total parameters and is designed for complex coding, visual reasoning, and professional work.
Alibaba Cloud lists a 1,000,000-token context window, with output limits of 131,072 tokens in non-thinking mode and 262,144 in thinking mode. The TokenHub model page shows its configured maximum output.
TokenHub lists text, images, PDFs, and video as inputs, with text output. The accepted attachment formats depend on the endpoint.
Alibaba Cloud supports low, high, and xhigh reasoning effort, with high as the default. The parameter name depends on the API protocol. Deeper reasoning can increase latency and token use.
Use your TokenHub API key and set the model field to qwen4-max. The API Integration section provides request examples for Chat Completions, Responses, and Messages.
The QwenCloud Singapore model page lists $1.50 per million input tokens, $4.50 per million output tokens, and $0.075 per million tokens for implicit cache hits. See this page’s Pricing table for TokenHub rates.
Choose Qwen4-Max when a task combines multiple steps with substantial visual or document context, such as implementing a design or reviewing a collection of reports. For demanding runs, account for the extra time and tokens needed for deeper reasoning.
Use one API key to access Qwen4-Max and more AI models through TokenHub.
Qwen4-Max In The Community
Launch posts, hands-on demos, and developer discussions about Qwen4-Max.
X (Twitter)
Qwen ·
📢 Meet Qwen4-Max, your new AI teammate for work, creativity, and discovery! Qwen4-Max is built on a new MoE architecture with native vision and a 1M-token context window. Highlights: - Outperforms Qwen3.8-Max with up to ~50% fewer tokens, while average cost per task drops over https://t.co/2Rlkoc6Dm3
https://pbs.twimg.com/media/HXNgxQIaEAAHHwZ.jpgView originalArtificial Analysis ·
Alibaba has launched Qwen4 Max, gaining 9 points over Qwen3.8 Max on the Artificial Analysis Intelligence Index and landing 3 points behind Kimi K3.1 Pro at ~11% lower Cost per Task Qwen4 Max, using xhigh reasoning effort, scores 43 on the Intelligence Index, 9 points ahead of https://t.co/dwh719Cq14
https://pbs.twimg.com/media/HXNuLl6akAA2GOq.jpgView originalQwen ·
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 https://t.co/3itCIgVAeE
https://pbs.twimg.com/media/HWzv8WRbEAA6avf.jpgView originalReddit
Qwen4-Max is officialy live!
forever_365 ·
View originalQwen4-Max is here. How competitive is it?
zyyyyl9144 ·
View originalQwen4-Max is out, and alibaba has cooked this time. Nearly 2.4T model, It's the combination of 71B active parameters.
Sea_Emergency_8458 ·
View originalYouTube
Qwen4-Max is now available.
Qwen
View originalQwen4-Max for 3D understanding and design.
Qwen
View originalQwen4-Max 正式版发布,9大场景编程测试,到底有多强
AI随风
View original