Microsoft launches MAI-Thinking-1, reducing OpenAI dependency

Microsoft has introduced MAI-Thinking-1 and a wider family of in-house AI models as it strategically reduces dependence on OpenAI to build a broader AI stack. The move comes as Microsoft expands its AI capabilities beyond just GPT integration, developing proprietary reasoning models that can handle complex enterprise workflows independently. MAI-Thinking-1 is designed specifically for enterprise decision-making and logical reasoning tasks, positioning Microsoft to compete more directly with Anthropic's Claude Opus and Google's Gemini models. This internal development aligns with Microsoft's broader strategy of creating end-to-end AI solutions for enterprise customers, from infrastructure to applications, rather than relying solely on external partnerships. The company has been investing heavily in its Azure AI infrastructure and is building more of its AI stack in-house.