"On Thursday, the company's in-house team announced MAI-Image-2, a second-generation image model that has debuted at number three on the Arena.ai text-to-image leaderboard, placing Microsoft's own technology directly behind Google's Gemini 3.1 Flash and OpenAI's GPT Image 1.5."
"MAI-Image-2 extends that trajectory: built with input from photographers, designers, and visual storytellers, and focused on three areas where creatives said the gap was largest. The first is photorealism, natural light, accurate skin tones, environments with physical texture and wear. Microsoft says the model is designed to reduce the post-production work that currently sits between generation and usable output."
"The announcement comes from the Microsoft AI Superintelligence team, the internal research group that Mustafa Suleyman formed in November 2025 and now leads full-time following a leadership reorganisation at Microsoft announced just two days ago."
Microsoft's AI Superintelligence team has released MAI-Image-2, a second-generation in-house image model that achieves third place on Arena.ai's text-to-image leaderboard, surpassing previous reliance on OpenAI's models. The model was developed with input from photographers, designers, and visual storytellers, focusing on three key improvements: photorealism with natural lighting and accurate skin tones, in-image text rendering for readable lettering and typography, and additional capabilities. This release follows Mustafa Suleyman's transition to lead the AI Superintelligence team full-time after a Microsoft leadership reorganization. MAI-Image-2 builds on MAI-Image-1, launched in October 2025, and integrates alongside DALL-E 3 and GPT-4o in Bing Image Creator and Copilot.
#image-generation-models #microsoft-ai-development #generative-ai-leaderboards #photorealism-and-text-rendering
Read at TNW | Microsoft
Unable to calculate read time
Collection
[
|
...
]