Microsoft Launches In-House AI Models Cutting Costs 89%

Microsoft AI has launched two new in-house models, MAI-Image-2.5-Pro and MAI-Voice-2-Flash, with a bold claim: they can deliver up to 89% cost savings compared to comparable offerings from OpenAI. The MAI-Image-2.5-Pro model is designed for high-fidelity image generation, targeting creative professionals and enterprises that require photorealistic visuals for marketing, design, and media production. On the other hand, MAI-Voice-2-Flash is optimized for high-volume enterprise voice workloads, such as call center automation, voice assistants, and real-time transcription services. By building these models internally, Microsoft aims to reduce dependency on third-party providers while offering competitive performance at a fraction of the cost. The 89% savings figure is based on inference costs, which are a major expense for businesses deploying AI at scale. This move is part of Microsoft's broader strategy to vertically integrate AI capabilities across its cloud and productivity ecosystem. For enterprises already using Azure, these models could significantly lower the total cost of ownership for AI-powered applications, making advanced image and voice AI more accessible to a wider range of businesses.