Building a hill-climbing machine: Launching seven new MAI models | Microsoft AI

Microsoft AI announced seven new models spanning image, voice, transcription, coding, and reasoning capabilities.

MAI-Thinking-1 is positioned as a flagship reasoning model matching leading models on software engineering benchmarks; MAI-Code-1-Flash is an inference-efficient coding model with 5 billion active parameters integrated into GitHub Copilot; MAI-Image-2.5 supports text-to-image and image editing; MAI Transcribe-1.5 is described as the best transcription model with five times faster performance than competitors across 43 languages; MAI-Voice-2 provides speech generation across 15 languages with voice adaptation capabilities.

The models will be available on Foundry, OpenRouter, Fireworks, and Baseten, with developers able to tune model weights.

Microsoft also announced a collaboration with Mayo Clinic to co-create a frontier AI model for healthcare.