Qwen4-Max Released with MoE Architecture and 1M Context Window
·2026.11.16 11:13
Key point
The new model is API-only and uses a MoE architecture, while smaller Qwen4 versions are confirmed for future open-weight release.
Details
- Qwen4-Max is officially released as an API-only model, with no indication that its weights will be made public.
- The model features a new MoE architecture, native vision capabilities, and a 1M-token context window.
- Internal evaluations show Qwen4-Max outperforms Qwen3.8-Max with up to ~50% fewer tokens and reduces average cost per task by over 60%.
- Smaller Qwen4 models are confirmed to be released as open-weight soon, with benchmarks for the Max version providing a basis for estimating their performance.
- The model targets real-world productivity tasks including 3D modeling, game development, manufacturing, finance, law, healthcare, and scientific discovery.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.