Microsoft on June 2 introduced MAI-Code-1-Flash, a coding model, and MAI-Thinking-1, a reasoning model, at its Build developer conference in San Francisco.

The move expands its in-house AI models as it seeks to compete more directly with partners including OpenAI and Anthropic.

Microsoft said MAI-Thinking-1 is a medium-sized model matching leading models on software engineering benchmarks and trained without distillation from third-party models. It said MAI-Code-1-Flash is designed for inference efficiency.

Microsoft also announced updated models for speech recognition, synthetic voice generation, and image generation, along with small Aion models for Windows PCs.

MAI-Code-1-Flash is integrated into GitHub Copilot and Visual Studio Code. MAI-Thinking-1 is in private preview through Microsoft Foundry, an Azure AI platform that consolidates services and provides model access through an Azure account and portal.

The launch follows Microsoft’s move into in-house models last year and a 2025 change to its OpenAI deal that removed exclusivity over OpenAI cloud computing, giving Microsoft more room to develop competing models.

🔗 Source: CNBC

Recent Microsoft developments