Microsoft Announces Seven MAI Models for Reasoning, Coding, Images, and Audio
At Build 2026, Microsoft announced MAI-Thinking-1, MAI-Code-1-Flash, MAI-Image-2.5, MAI-Image-2.5-Flash, MAI-Voice-2, MAI-Voice-2-Flash, and MAI-Transcribe-1.5. Microsoft says MAI-Thinking-1 was trained without OpenAI data and matches Claude Opus 4.6 on SWE Bench Pro, with 35B active parameters and a 256K-token context. The lineup also covers transcription in 43 languages, emotion-controlled speech generation in 15 languages, and image editing. Microsoft announced proprietary distribution through Microsoft Foundry and GitHub Copilot.
Visit original source