Alibaba launches Qwen4-Max, a closed 2.4-trillion-parameter flagship model
Alibaba has made Qwen4-Max, its new flagship model, generally available through QwenCloud after a limited preview that began on November 11, TechNode reports. The model is built on the new Qwen4 architecture with a 2.4-trillion-parameter mixture-of-experts backbone, 71 billion parameters activated per token, native vision and a one-million-token context window, according to the Qwen blog.
Unlike some earlier Qwen releases, Qwen4-Max is closed: Alibaba did not release its weights, though it says smaller Qwen4 models will be open-sourced, without giving a date. In Alibaba's own coding tests, Qwen4-Max scored 75.7 on DeepSWE v1.1, against 74.2 for Anthropic's Opus 5.5 and 75.2 for OpenAI's GPT-6.1 Sol, India Today reports.
Alibaba also says the model is cheaper to run. On Work360, its in-house benchmark of work tasks, the average cost per task fell from $4.43 with Qwen3.8-Max to $1.67, because the new model uses fewer tokens and costs less per token, MLQ.ai reports. For cybersecurity, Alibaba says it kept training data for defensive work and removed data meant to strengthen attack capabilities.
We hope you enjoyed this article
Consider subscribing to one of our newsletters like Daily AI Brief.
Also, consider following us on social media:
Daily AI Brief
Daily report covering major AI developments and industry news, with both top stories and complete market updates
Market report
2025 Generative AI in Professional Services Report
Thomson Reuters
This report by Thomson Reuters explores the integration and impact of generative AI technologies, such as ChatGPT and Microsoft Copilot, within the professional services sector. It highlights the growing adoption of GenAI tools across industries like legal, tax, accounting, and government, and discusses the challenges and opportunities these technologies present. The report also examines professionals' perceptions of GenAI and the need for strategic integration to maximize its value.
Read moreYou may also like
MIT researchers use GPT-6 Astra to edit Qwen model weights
Nous Research releases Arete-27B, a Qwen model post-trained to remove CCP-aligned answers
Cohere Says Alibaba's Qwen Model Tried to Send Data to Beijing Server
Chinese Models Underpin Nearly Half of New Korean AI Releases
New coding benchmark puts Claude Opus 5.5 on top and opens a 35-point gap that DeepSWE hid
Daily AI Brief: the AI news that matters, in your inbox.