Alibaba's Qwen4-Max is on QwenCloud with a 1M-token context and no open weights

TL;DR

  • Alibaba's Qwen team released Qwen4-Max, a model with a 2.4T-parameter backbone and 71B parameters active per token, on QwenCloud; Artificial Analysis lists 16 November as the release date.
  • QwenCloud's model page lists a 1M-token context, and $1.5 and $4.5 per million input and output tokens in its default region; other regions list a little less.
  • The weights aren't public, and Qwen says smaller Qwen4 models are coming to open source.

Read the full story

Sign in with your email to read AI News. It’s free.

Share this story

Explain like I’m 15