GPT-5.6 is now available in Figma Make

With accurate first results and fast iterations, OpenAI’s GPT-5.6 gives your builds a strong start in Figma Make.
Share GPT-5.6 is now available in Figma Make
From a rough sketch to a polished prototype, or even production code, teams can use Figma Make to move faster from idea to execution. Today we’re adding GPT-5.6—OpenAI’s newest model—to Make. From our early tests, GPT-5.6 raises both the quality and speed of what Make produces on a first pass. From clean UI to responsive layouts to high fidelity interactions, you can get quality prototypes, fast.
Build fast so you stay in your flow
Going from a prompt to a working build quickly makes it easy to explore and validate more directions in a single sitting. Our tests have shown that GPT-5.6 does this capably, even for complex designs. For example, one of the evals we run regularly tests the model’s ability to go from 0-1 in creating a stock tracking app with a dark, gothic aesthetic. A recent run produced an interactive dashboard with sample price values, performance figures, and new details like keyboard shortcuts and search—all within a single prompt and with greater token-efficiency While everyone’s talking about “tokenmaxxing,” we’ve compiled seven best practices to help you build smarter—without prompting more—in Figma Make.
7 tips for using Figma Make credits more efficiently
GPT-5.6 also keeps moving when a build hits an error. Instead of stopping, it investigates and self-heals, which means less back-and-forth to reach a working result. In one case, GPT-5.6 identified the source of a blank build and fixed it on its own. Getting to a solid prototype sooner means you can stay in the flow and bring your team in earlier to react, refine, and build on ideas together.
Keep prototypes true to existing designs
GPT-5.6 excels at building from an existing design, like a design spec or Figma Design file. In one test, we started prompting from a Figma Design file for a nature sound player app with playable tracks and music controls. The design we provided included a multi-track timeline of layered sound clips, playback controls, and a sound library with artwork. The first pass came back faithful to the original design. The layout and visual hierarchy, spacing and proportions, and styling all carried over.
The interactions worked well, too. On the sound player, the play, pause, and skip controls all functioned as expected, and the audio files played back correctly. GPT-5.6 makes building an interactive prototype from a static design a dependable step in the design-to-code workflow.
Produce high-quality first passes
A usable prototype needs polished styling, interactions that work, and a layout that adapts across screen sizes. We gave GPT-5.6 a minimal prompt to build a content-heavy e-commerce website with a bookshelf product catalog. The first pass came back clean and professional, and it included product descriptions, measurements, and care guides.
All the interactions also worked right off the bat, including dropdowns with populated product descriptions, an interactive photo library for the bookshelf, and a clickable menu. Even without any additional prompting, the build adapted across screen sizes reliably.
When a prototype starts closer to the vision, it’s a launch pad for teams to build on, together. That’s what GPT-5.6 brings to Make: a faster way to get from idea to execution, without losing craft along the way.
GPT-5.6 is available in Figma Make today. Head to the model selector in Make and click on GPT-5.6 to try it out.
Learn more about selecting and working with this new model in our help center.

Gui is the Design Director for AI at Figma, leading work on agentic workflows for design and code. Previously, he led AI initiatives at Pinterest.



