Introducing MAI-Image-1: Microsoft’s In-House Image Generation Model in Bing Image Creator and Copilot Audio Expressions
stclarke introduces MAI-Image-1, a Microsoft-developed AI model for image generation, now available in Bing Image Creator and Copilot Audio Expressions, and discusses its capabilities and integration into Microsoft products.
Introducing MAI-Image-1: Microsoft’s In-House Image Generation Model
Microsoft has unveiled MAI-Image-1, its first image generation model developed entirely in-house, now available to creators through Bing Image Creator and Copilot Audio Expressions. This new model not only entered the top 10 in the LMArena text-to-image leaderboard but also showcases Microsoft’s ongoing commitment to building innovative, user-centric AI technologies.
Where to Try MAI-Image-1
- Bing Image Creator: Find MAI-Image-1 at bing.com/create, in the Bing mobile app, or directly in Bing’s search bar. Users can choose MAI-Image-1 alongside DALL-E 3 and GPT-4o to fit their creative needs.
- Copilot Audio Expressions: Selecting Story Mode in Audio Expressions (link) now uses MAI-Image-1 to visualize your stories with custom-generated images.
Note: MAI-Image-1 is available in all countries where Bing Image Creator and Copilot Labs are accessible, except in the EU.
Model Highlights and Capabilities
- Photorealistic Imagery: Excels in rendering intricate lighting, landscapes, natural scenes, and more, performing competitively against larger, slower models.
- Efficiency: Offers a balance of speed and visual quality, enabling users to quickly iterate and refine their ideas.
- Flexibility and Diversity: Designed to avoid repetitive or generic results and deliver imagery that meets diverse creative demands.
Development Approach
- Purpose-Built for Creators: MAI-Image-1 was developed with an emphasis on real-world use cases, informed by feedback from creative professionals.
- Data Selection & Evaluation: The team prioritized rigorous dataset curation and nuanced evaluation methodologies to ensure flexibility and authenticity in generated images.
Microsoft’s Vision for AI
Microsoft describes this release as a step toward their mission to create AI that is supportive, helpful, and contributes positively on a global scale. The ongoing development of in-house models like MAI-Image-1 supports more immersive and useful AI-driven experiences across Microsoft’s product ecosystem.
Getting Involved
Microsoft AI highlights its ambition and active AI research and development teams. Interested professionals are encouraged to learn more about opportunities at MAI.
Further Reading
Author: stclarke
This post appeared first on “Microsoft News”. Read the entire article here