Microsoft revealed MAI-Thinking-1, its first advanced reasoning artificial intelligence model, during its Build 2026 conference in May 2026 [1]. The model features 35 billion active parameters, placing it among medium-sized but powerful AI systems [2].
MAI-Thinking-1 was trained from scratch using clean data and does not rely on any third-party models, differentiating it from many other AI offerings [1]. Microsoft stated that MAI-Thinking-1 matches leading AI models on key software engineering benchmarks, highlighting its strong performance in coding and reasoning tasks [1].
Alongside MAI-Thinking-1, Microsoft announced six additional AI models targeting fields such as image generation, transcription, voice recognition, and coding assistance [1]. These launches demonstrate Microsoft's broad investment in AI advancements beyond language reasoning.
Microsoft’s announcement in May 2026 marked the company’s latest step into advanced AI technology, aiming to compete with leading AI developers by offering new capabilities optimized for software engineering and multimodal AI applications [1, 2].
