Microsoft Releases MAI-Image-2.6-Flash with 2x Faster Image Generation Speed
Microsoft has released MAI-Image-2.6-Flash in public preview on Microsoft Foundry, offering twice the generation speed of models like GPT-Image-2-Medium. It delivers 72% higher GPU efficiency while cutting image output API costs in half compared to its flagship MAI-Image-2.6 model. The release significantly lowers latency and operational expenses for high-volume AI image generation and visual editing workflows. This makes high-throughput image applications more practical and affordable for developers and enterprise clients. MAI-Image-2.6-Flash retains core capabilities of the base model, including object replacement, inpainting, attribute modification, and text editing. Microsoft recommends using the flagship MAI-Image-2.6 for high-quality production assets, while directing latency- and cost-sensitive tasks to the Flash variant.
## BACKGROUND
MAI-Image models are Microsoft's proprietary diffusion-based generative AI models designed for text-to-image creation and controllable image editing. Microsoft Foundry, formerly known as Azure AI Studio, serves as the platform where developers explore, evaluate, and deploy these generative AI models.