Key Info

Qwen4-Max is now officially released, built on a new MoE architecture with native vision and a 1M-token context window. The team says it outperforms Qwen3.8-Max with up to ~50% fewer tokens, while average cost per task drops over 60% in internal testing.

Highlights

  • New MoE architecture with native vision support
  • 1M-token context window
  • Up to ~50% fewer tokens than Qwen3.8-Max on comparable tasks
  • Average cost per task down over 60% in internal evaluations
  • More Qwen4 models are stated to be on the way