Microsoft unveils MAI-Image-2, an ambitious AI image generator

⚡

Key Takeaways

1Microsoft, under the leadership of Mustafa Suleyman, has launched MAI-Image-2, an AI image generator that ranks third on Arena.ai.

2MAI-Image-2 excels at creating photorealistic images and rendering text for various uses such as posters.

3The model is available for testing in MAI Playground and will be integrated into Copilot and Bing Image Creator.

💡Why it matters — Microsoft is looking to catch up with OpenAI and Google in the rapidly growing field of AI image generators.

⚡Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisis ton rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

📄

Full Analysis

Microsoft Advances in the AI Race with MAI-Image-2

Microsoft's superintelligence team, led by Mustafa Suleyman, has recently launched MAI-Image-2, an artificial intelligence model designed to transform textual descriptions into images. This model currently ranks third in Arena.ai's leaderboard for text-to-image generators, behind OpenAI's GPT-Image-1.5 and Google's Nano Banana 2, with a significant gap.

MAI-Image-2 stands out for its ability to produce particularly realistic images, featuring natural lighting and accurate skin tones. Microsoft collaborated with photographers, designers, and visual artists to develop this model, which is capable of handling detailed and surreal scenes.

Photorealistic Images and Reliable Text Rendering

The MAI-Image-2 model generates photorealistic images with fine details, such as portraits with shadow play, macro shots of irises, and scenes of glacial caves. Additionally, it excels in practical tasks such as reliably rendering text within images, making it useful for creating posters, infographics, and diagrams.

Availability and Future Prospects

MAI-Image-2 is currently available for testing in the MAI Playground, depending on the region, and will be integrated into Copilot and Bing Image Creator. For now, API access is limited to select enterprise clients, but Microsoft plans to open it up to all developers via Microsoft Foundry. No detailed technical information, pricing, or training data has been shared by the company.

The launch of MAI-Image-2 marks a significant advancement over its predecessor, MAI-Image-1, which was released in October 2025 and ranked ninth on the same leaderboard, failing to generate much interest. Microsoft aims to continue progressing to compete with industry leaders, OpenAI and Google.

⚡

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.

📰 Voir toutes les actus IA →