Alibaba Qwen3.8-Max: 2.4T MoE with Max-class open weights next week
Alibaba released Qwen3.8-Max, the most capable Qwen model to date—a 2.4-trillion-parameter mixture-of-experts that activates about 95B parameters per request, with a context window up to 1 million tokens. Built on the Qwen 3.5 architecture, it targets coding, real-world work, research, and long-horizon agent tasks, and Alibaba says it is the first Max-class Qwen model it will open-source (weights planned next week on Hugging Face and ModelScope). The model is available now via QwenCloud / Alibaba Cloud Model Studio APIs, including OpenAI- and Anthropic-compatible endpoints for coding agents.






