Microsoft is expanding the use of its own AI models across various products, replacing some reliance on OpenAI's GPT models. The company has fully transitioned the Bing Image Creator to its MAI-Image-2.5 model, which offers improved performance in image generation and editing compared to GPT-Image-2. This includes features like removing objects, enhancing specific areas, and reimagining images through natural language commands. In PowerPoint, MAI-Image-2.5 enables image-to-image transformations with reduced GPU costs by up to 84%. OneDrive now uses MAI-Image-2.5 as the default model for image editing, offering faster response times and increased storage efficiency. Additionally, Microsoft introduced MAI-Image-2.5-Pro for high-quality tasks where accuracy is critical, such as precise text rendering in generated images. The company also launched MAI-Voice-2-Flash for speech-related applications like customer service and Azure services.
Bias read (Center): The article presents a factual update on Microsoft's strategic shift toward using its own AI models, focusing on technical improvements and cost efficiencies. It does not take a clear ideological stance, nor does it emphasize political implications or controversies. The framing remains neutral, with





