Key Info

Qoder now offers Day-0 support for Qwen4-Max, a new MoE-based model with native vision and a 1M-token context window, positioned for long-horizon agentic coding tasks.

Highlights

  • Qwen4-Max is built on a new MoE architecture with native vision support.
  • It offers a 1M-token context window for handling large tasks.
  • The model reportedly outperforms Qwen3.8-Max while using up to ~50% fewer tokens.
  • Average cost per task drops over 60% in internal testing, per the announcement.
  • Users can switch models inside Qoder and hand it larger tasks directly.