Microsoft Releases MAI-Image-2.6, Surges to No. 2 on Arena

Microsoft has introduced MAI-Image-2.6, the latest version of its in-house AI image generation model, with improvements to image quality, text rendering, and reference image support.

The launch comes shortly after Microsoft released MAI-Image-2.5-Pro. According to Microsoft, MAI-Image-2.6 has already climbed to the No. 2 position on the Arena leaderboard.

The model currently trails only OpenAI’s GPT-Image-2. Microsoft says MAI-Image-2.6 scored 79 Elo points higher than MAI-Image-2.5 in Arena’s text-to-image category.

MAI-Image-2.6 improves text, portraits, and 3D images

Microsoft says the new model delivers better text rendering, portraits, and 3D imagery compared with its previous image generation models.

MAI-Image-2.6 also focuses on producing more polished commercial and photorealistic results. Microsoft says the model performs particularly well for product imagery, branding, and cinematic content.

The company has also improved grounding and support for workflows that use multiple reference images. Users get more control over reasoning, output format, and image resolution.

MAI-Image-2.6 is already available through Arena

Users can already test MAI-Image-2.6 for image generation through Arena.

Microsoft plans to add the model to MAI Playground later this week. The company also intends to roll it out through Microsoft Foundry and other Microsoft products and services.

Microsoft says it will share more details about MAI-Image-2.6 in the coming weeks.

In other AI news, OpenAI has launched GPT-5.6-Cyber as it expands its Daybreak cybersecurity program.