Alibaba launches Qwen4-Max with a million-token context and tool support

Alibaba launches Qwen4-Max with a million-token context and tool support

Ondřej Barták
Ondřej Barták
Entrepreneur and Programmer
16. 11. 2026
2 minutes reading · 1 views
Alibaba launches Qwen4-Max with a million-token context and tool support

Alibaba has launched Qwen4-Max, a new generation of its proprietary flagship model. It accepts text, images and video, has a one-million-token context window and supports tool use. The model is available through a cloud API and Alibaba applications.

Long context and tool use

Qwen4-Max is built on a redesigned mixture-of-experts architecture.

Alongside its multimodal input, the model supports tool calling, structured output and web search. Its output can reach 393,216 tokens, a limit shared between the model’s reasoning and its final response.

Alibaba is targeting multistep tasks that require long context and repeated tool use. Its intended applications include 3D design, game development and converting manufacturing sketches into production specifications.

How it compares with Qwen3.8-Max in Alibaba’s tests

In Alibaba’s internal evaluations, the new model matched or exceeded Qwen3.8-Max while using up to 50% fewer tokens. Qwen4-Max scored 52.69 on the company’s Work360 benchmark, which Alibaba says is roughly 10% higher than its predecessor.

Alibaba created Work360 to evaluate tool use and long-running professional tasks. It covers 26 main domains, 182 subdomains and 721 specialized fields.

In Work360, the company reports approximately 24% lower token consumption at the Xhigh reasoning level and 35% lower consumption at High. In Terminal-Bench 4.0, its tests showed a reduction of approximately 48% to 49%.

Access through APIs and applications

The model is available through QwenCloud API, Qwen Studio, QwenWork and the Qoder development environment. QwenCloud offers interfaces compatible with OpenAI Chat Completions, OpenAI Responses and Anthropic Messages.

Migration of an existing client begins by changing the API key, base URL and model identifier.

Qwen4-Max costs $1.50 per million input tokens, compared with $2 for Qwen3.8-Max; output costs $4.50 per million tokens instead of $6, and implicitly cached tokens cost $0.075 per million instead of $0.25.

Qwen4-Max remains proprietary and available through hosted services. The announced family also includes Qwen4-Flash, Qwen4-Plus and Qwen4-27B; Alibaba promises to release the weights of smaller Qwen4 models but has not set a date.

Advertisement

Content created with help from UpTier.

SEO and GEO on autopilot. UpTier’s multi-agent systems write and optimize content for search engines and AI answers.

Discover UpTier ↗

Category:AI
Did you enjoy this article?
Discover more interesting posts on our blog
Back to blog

Related posts

OpenAI introduces GPT-6.2 Astra and Sol for professional and everyday workOpenAI introduces GPT-6.2 Astra and Sol for professional and everyday work
OpenAI has introduced two models: Astra for complex professional tasks and Sol for everyday work. Both are available in ChatGPT Work, Codex and through the API.
3 min read
17. 11. 2026
Tencent releases Hy4 Flash with image support and a context window of up to 1 million tokensTencent releases Hy4 Flash with image support and a context window of up to 1 million tokens
The lighter Hy4 Flash model targets chat, documents and routine code edits. It accepts text and images, supports tool use and, according to Tencent, matched Claude Haiku 5.5 in tests of everyday tasks.
1 min read
16. 11. 2026
Companies Linked to Chinese Spy Invested in DeepSeek and Promised Banned Nvidia ChipsCompanies Linked to Chinese Spy Invested in DeepSeek and Promised Banned Nvidia Chips
Companies linked to an alleged Chinese spy joined DeepSeek’s latest funding round. In exchange for a stake, they reportedly promised banned Nvidia Blackwell chips.
3 min read
16. 11. 2026
Přihlaste se k odběru našeho newsletteru
Zůstaňte informováni o nejnovějších příspěvcích, exkluzivních nabídkách, a aktualizacích.
CodedTrip

Operated by CodedTrip LLC, USA.

YouTube
TikTok