415.tech
AI & tech, from the frontlines of Silicon Valley
Microsoft's MAI-Thinking-1 matches Claude Opus 4.6 on coding — trained without OpenAI data

Microsoft's MAI-Thinking-1 matches Claude Opus 4.6 on coding — trained without OpenAI data

Microsoft released MAI-Thinking-1, a 35B-active-parameter reasoning model trained from scratch on commercially licensed data with zero OpenAI distillation, scoring 97.0% on AIME 2025 and matching Claude Opus 4.6 on SWE Bench Pro coding benchmarks. Microsoft now has a competitive frontier reasoning model on its own Foundry platform independent of the OpenAI partnership, with MAI-Code-1-Flash rolling out immediately to all GitHub Copilot subscribers.

Source: blogs.microsoft.com ↗

Post on XEmail

Trained from scratch with zero distillation on enterprise grade, clean and commercially licensed data you can build on with confidence.

Microsoft AI Superintelligence Team

Why this matters

  • → Microsoft now has a competitive frontier reasoning model independent of OpenAI, reducing reliance on external partnerships for core AI capabilities.
  • → MAI-Thinking-1 matches Claude Opus 4.6 on coding benchmarks while trained exclusively on licensed data with no distillation.
  • → The model's 35B active parameters and 256K context window deliver high performance at lower token costs.
Building without OpenAI
Also in this edition