Figma Tests Nano Banana 2 for Faster AI Image Generation.

Figma has begun testing a new AI model called Nano Banana 2 that promises faster image generation and higher-quality outputs across its design platform.

The company announced on LinkedIn that the upgrade is rolling out to both Figma and Figma Weave, its AI-powered design assistant. The brief post highlighted two key improvements: faster outputs and "pro-level image generation," suggesting the model delivers both performance gains and a quality bump over its predecessor.

The timing follows Figma's aggressive push into AI-assisted design. The company recently opened beta access for AI agents that can design directly on canvas, and has been steadily expanding integrations with AI code editors like Cursor and Windsurf through its MCP server. Nano Banana 2 appears to be the next evolution in that strategy, focusing specifically on generative image capabilities.

While Figma hasn't disclosed technical details about the model architecture or training, the "Nano" naming suggests a focus on efficiency. This aligns with broader industry trends toward smaller, faster AI models that can run more responsively in production environments without sacrificing output quality.

The mention of Figma Weave is notable. Weave has become central to how Figma surfaces AI features to users, handling everything from auto-layout suggestions to content generation. Adding a faster, more capable image model to Weave could significantly reduce latency for designers generating mockups, illustrations, or placeholder assets during the design process.

Figma's claim of "pro-level" quality raises the bar for in-tool image generation. Many design platforms have added AI image features, but quality has been inconsistent, often requiring designers to use standalone tools like Midjourney or other specialized generators when they need publication-ready assets.

If Nano Banana 2 delivers on that promise, it could keep more workflows inside Figma, reducing context switching and speeding up iteration cycles. For teams already using Figma for prototyping and design systems work, having high-quality generative imagery available natively would remove a common friction point.

The LinkedIn post doesn't specify whether Nano Banana 2 is available to all users or limited to a subset during testing. Figma has historically rolled out AI features gradually, often starting with enterprise customers or beta programs before wider release. Users interested in trying the new model should check their Figma Weave settings or watch for in-app notifications.

For builders exploring AI-assisted design workflows, this update is worth watching. As tools like Lovable, v0, and Bolt compete to streamline design-to-code pipelines, Figma's investment in faster, smarter generation keeps the platform competitive for both traditional designers and those experimenting with vibe coding approaches.

Figma has not yet published a formal blog post or documentation detailing Nano Banana 2's capabilities, pricing implications, or rollout timeline. The company's official blog remains the best source for updates as testing progresses.