OpenAI launches GPT-4o Visual v2 with native Figma plugin for designers

AI · 5 min read

OpenAI announced GPT-4o Visual v2, a multimodal model optimized for creative workflows and desktop responsiveness. The release emphasizes lower latency on visual inputs and improved alignment for design tasks such as layout generation, asset recoloring, and contextual copy suggestions.

Alongside the model, OpenAI published a native Figma plugin that connects directly to GPT-4o Visual v2 through a new OAuth-based integration. The plugin can generate multiple wireframe variations, propose accessible color palettes based on WCAG checks, and inline suggested microcopy tied to selected frames or components.

Designers testing the plugin reported significant speedups for ideation and iteration, though some flagged occasional stylistic drift on brand-specific language. OpenAI said teams can fine-tune behavior using workspace-level prompts and that enterprise customers will get on-premises options later this year.