Google has unveiled Nano Banana 2 Lite, a new image generation model within the Gemini family, targeting speed and cost-effectiveness rather than just high visual quality. The idea is that companies and developers don’t want one stunning image; they want thousands of quick attempts to design an ad, experiment with an interface, or build editable visual content. Google is pushing for a model capable of producing images in seconds, later linked to a video model to turn static images into moving scenes.

Image in 4 Seconds

According to an announcement posted on the Google blog, Nano Banana 2 Lite is Gemini’s fastest and most cost-effective image model. It generates text-to-image outputs in 4 seconds at $0.034 per 1,000 1K images. It is available through Google AI Studio, the Gemini API, and the Gemini Enterprise platform. It will also be rolled out within consumer products like AI Mode in Search, the Gemini app, and other products.

Image and Video in One Line

Google has also launched Gemini OmniFlash for developers, a model for video generation and editing through conversation. Developers can use Nano Banana 2 Lite to produce an image quickly and then pass it to OmniFlash to turn it into a video. The models use the SynthID watermark to verify AI-generated content, placing the news in a sensitive area between accelerating visual production and concerns about misinformation and synthetic content spreading on platforms.