Google unveils Nano Banana 2 Lite, previews Gemini Omni Flash for developers

Google has unveiled Nano Banana 2 Lite and announced the preview availability of Gemini Omni Flash, expanding its suite of AI-powered media generation models for developers and NotebookLM users. The company says Nano Banana 2 Lite is its fastest and most cost-effective image generation and editing model to date.

Compared with the original Nano Banana model, Nano Banana 2 Lite is designed to deliver faster image generation while offering improved visual quality and enhanced capabilities.

Among its new features is the ability to generate context-aware scenes, location-specific mockups and preliminary data visualisations using real-world knowledge. The model also aims to maintain character and object consistency across multiple image generations, making it suitable for applications such as storyboarding and virtual try-on experiences for e-commerce.

Additionally, Nano Banana 2 Lite can render clear, legible text within images, enabling users to preview typography and create localised visual assets for different languages and markets.

More from Tech

Google has begun rolling out Gemini Omni Flash to developers through the Gemini API and Google AI Studio. The model natively supports high-quality video generation and conversational editing using a combination of text, image and video inputs. Google has priced the model at $0.10 per second of video output, matching the cost of Veo 3.1 Fast.

Streamlined video editing with conversational prompts

Gemini Omni Flash introduces conversational video editing, allowing users to modify videos using natural language prompts. It also supports multimodal referencing, enabling users to combine text, images and video inputs to maintain greater control and consistency across generated scenes.

The model leverages Gemini's world knowledge—including history, biology and narrative reasoning—to generate contextually relevant videos. Users can also add text overlays and graphics directly into videos through simple prompts.

At launch, Gemini Omni Flash supports video generation of up to 10 seconds, with longer durations expected in a future update. The current version does not support audio reference uploads or scene extension through the Gemini API. Google has also noted that the model has limitations when handling panning camera movements.

Alongside the rollout, Google introduced a set of demo applications showcasing the combined capabilities of Nano Banana 2 Lite and Gemini Omni Flash. The Anywhere app lets users upload a selfie or photo and place themselves at iconic landmarks before generating an animated version of the scene. Space Lift is an AI-powered interior design tool that transforms room photos into redesigned spaces across different design styles and creates animated visualisations of the concepts. Meanwhile, Omni Product Studio converts static images into cinematic product videos, demonstrating AI-powered image-to-video generation for e-commerce content.