Ant Group open-sources six billion design models for UI posters and layer editing

ME News report, September 23 (UTC+8), Beating AI Bulletin: Ant Group’s InclusionAI has open-sourced the Ming-Image-0.1-Design series, comprising two 6B models. The Design model specializes in generating UIs, dashboards, infographics, and posters; the Layer model decomposes an already-generated design into individual transparent layers that can be moved, replaced, or modified independently. Both models are licensed under MIT. Design supports structured prompts up to 8K tokens, enabling it to generate entire pages in one go and directly produce RGBA assets with transparent backgrounds. Layer can split designs into 2 to 9 layers—including text, cards, main elements, and background—and reassemble them back into the original image. Ant Group has also open-sourced Design Skill and PPT Skill. The former integrates design generation, asset decomposition, and front-end coding; the latter reconstructs a flat image into a PowerPoint file that allows further editing of text, images, and layout. In Ant Group’s publicly released Artificial Analysis UI/UX Design open-weight leaderboard, Ming-Image-0.1-Design ranks first with an Elo score of 1082. The models are now available on OpenRouter for free access. (Source: BlockBeats)