Microsoft releases MAI Code 2 and Code 2 Flash for enterprise coding

·2026.11.18 03:15

Key point

The new models deliver 98% of GPT 6.1 Sol's performance at 40% of the cost and 96% of Claude Opus 5.5's performance at 5% of the cost.

1 / 2

Details

  • Microsoft introduced MAI Code 2 and MAI Code 2 Flash, two coding models optimized for everyday enterprise tasks like debugging, multi-file implementation, and UI generation.
  • MAI Code 2 achieves 98% of GPT 6.1 Sol's performance at 40% of the cost and 96% of Claude Opus 5.5's performance at 5% of the cost across four benchmarks.
  • MAI Code 2 Flash targets high-volume routine tasks, offering performance comparable to GPT-6 Luna and Claude Haiku 5.5 at 65% and 20% of their respective costs.
  • In blind frontend evaluations, reviewers preferred MAI Code 2's visual coding results over GPT-6 Sol 67% of the time.
  • The models are currently available in GitHub Copilot, with Code 2 Flash in limited-access preview on Microsoft Foundry and upcoming releases on Open Router, Vercel, Hermes, Open Code, and Kilo.
  • MAI Code 2 also powers Microsoft products, matching or exceeding GPT 6.1 Sol and GPT 6 Luna in accuracy and token efficiency for Excel, PowerPoint, Word, Teams, and Fabric IQ.

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.