Alibaba released Qwen4-Max, which it called the most capable model in the Qwen family, the company said on Monday, Nov. 16. In a statement posted on Alibaba Cloud, the company said the flagship is a mixture-of-experts model with 2.4 trillion total parameters that activates 71 billion parameters per token, a one-million-token context window, and native vision capabilities.
Bloomberg also reported the launch. Alibaba said Qwen4-Max is available now on QwenCloud through a token plan and an API, and on Alibaba platforms including QwenWork and QwenStudio. The company said open-weight, smaller-scale models from the Qwen4 series will soon be released to developers. Bloomberg reported that Alibaba did not make Qwen4-Max itself open-weight.
Alibaba said the new architecture upgrades a sparse mixture-of-experts design and activates 25% fewer parameters than Qwen3.8-Max at the same 2.4 trillion-parameter scale. On Work360, a proprietary benchmark the company said covers industry tasks across 721 specialized fields, Alibaba said Qwen4-Max outperforms Qwen3.8-Max overall while using 35% fewer tokens in a high-effort reasoning mode and reducing the average cost per task by around 60%. The company listed an input price of $0.075 per 1 million tokens for an “Implicit Cache Hit.”
Alibaba said Qwen4-Max “achieves top rankings across leading benchmarks, particularly in coding, agentic tasks and visual reasoning,” and that it ranked sixth on Code Arena, which the company described as a real-world coding leaderboard. Bloomberg reported that Alibaba said the model scored near the top of coding benchmarks, outperforming Anthropic’s and Moonshot’s latest models.
Artificial Analysis, which publishes an independent model index, lists Qwen4-Max at 43 on its Intelligence Index, 38th of 81 comparable models on that page (median 41), and shows a 984,000-token context window. The company release states a one-million-token window.
Alibaba also said it filtered cybersecurity training data to keep defensive material and remove offensive attack content, and that it added protections against chemical, biological, radiological, and nuclear risks. AI Tech Daily has previously reported on Alibaba’s Qwen models.
Sources
- Alibaba Cloud, Alibaba Unveils Qwen4-Max: New AI Teammate for Work, Creativity, and Discovery, Nov. 16, 2026
- Luz Ding, Bloomberg, via Yahoo Finance, Nov. 16, 2026
- Artificial Analysis, Qwen4 Max, accessed Nov. 16, 2026
- AI Tech Daily, Cohere says Alibaba Qwen model tried to upload a test file to a Beijing cloud bucket (October 27, 2026)