Google launches Nano Banana 2 Lite and Gemini Omni Flash

Google has announced two additions to its generative media model family: Nano Banana 2 Lite and Gemini Omni Flash. The releases suggest a continued push toward making image and video generation more accessible, whether through lower-cost API access or early previews of new video capabilities.

Nano Banana 2 Lite is positioned as a cost-efficient text-to-image model, available through the Gemini API and integrated into Google products including Search and Photos. The "Lite" designation typically signals a model optimized for speed and affordability rather than maximum output quality, making it a practical option for developers and applications where volume or cost is a concern rather than top-tier fidelity.

Gemini Omni Flash, meanwhile, is entering preview as a video generation model capable of producing up to 10-second clips from text prompts, still images, or existing video clips. The ability to accept multiple input types gives it some flexibility - users are not limited to text-only prompts, which opens up use cases like extending or stylizing existing footage. At 10 seconds, the output length is modest but consistent with where many short-form video generation models currently sit.

The dual release reflects Google's broader strategy of shipping models across a range of price points and capability tiers. With Nano Banana 2 Lite handling image generation at lower cost and Gemini Omni Flash targeting video - albeit still in preview - Google is filling out its generative media stack incrementally rather than waiting for a single large launch.