Alibaba launches Qwen4-Max: MoE with native vision, 1M context, 60% cheaper per task

danielhanchen · x · 2026-11-16

Qwen officially announced Qwen4-Max: a new MoE architecture with native vision and a 1M-token context window. Per internal evaluations it outperforms Qwen3.8-Max with 50% fewer tokens and cuts average cost per task by over 60%. Pricing: $1.5/M input, $4.5/M output, $0.075/M implicit caching. Alibaba also teased smaller Qwen4 models soon, with Qwen4-27B claimed to run on 16GB RAM setups.

Related event: Alibaba Releases Qwen4-Max with Native Vision and 1M Context(12 posts)→

Original post →

More from Models

Models channel →