Create Interfaces from a Prompt: Figma Gets Gemini 3 Pro
Users can provide a text prompt or a reference image, and the model produces interface structures, stylistic variations, and even generates HTML/CSS. It also supports interactive prototypes with animations, states, and transitions.
Google has introduced Gemini 3, its new flagship AI model combining enhanced contextual understanding, multimodal capabilities, and strong reasoning performance. One of the first platforms to receive practical integration is Figma: Gemini 3 Pro is now available in Figma Make as an experimental AI feature.
In Figma Make, Gemini 3 Pro generates layouts, components, and basic design systems. Users can provide a text prompt or a reference image, and the model produces interface structures, stylistic variations, and even generates HTML/CSS. It also supports interactive prototypes with animations, states, and transitions, speeding up the path from concept to functional prototype. Some manual refinement may still be required for complex or unconventional designs.
Gemini 3 Pro has already topped benchmark rankings, leading multiple categories in Design Arena-including web design and UI components-and performing strongly in LMArena. According to Figma, the integration lowers the entry barrier for newcomers and reduces routine workload for professionals, though it doesn’t replace design thinking or UX expertise.
Overall, the arrival of Gemini 3 in Figma signals a new level of interface-creation automation-from fast generation to code-ready prototypes-making design work quicker and more accessible.
Dylan Field, co-founder and CEO, showcases several early experiments with Gemini 3 Pro that we particularly liked.