Google DeepMind on June 30 released two new generative AI models designed to make image and video creation faster and cheaper for developers: Gemini Omni Flash for video generation and Nano Banana 2 Lite for rapid image output.
Conversational video editing
Gemini Omni Flash (gemini-omni-flash-preview) is a multimodal model that generates videos up to 10 seconds long at 720p from text descriptions, still images, or existing video clips. Its headline feature is conversational editing: instead of cutting a timeline, users describe changes in plain language and the model applies them, maintaining text-action synchronization and drawing on real-world context to render scenes accurately.
The model is priced at $0.10 per second of video output and is available now through Google AI Studio, the Gemini API, the Gemini app, and Google Flow. Google notes current limitations — no audio-reference support, a 10-second length cap, and uneven character consistency across scene changes — with longer videos and scene extension coming in future updates.
Fast, cheap image generation
Nano Banana 2 Lite (gemini-3.1-flash-lite-image) is now generally available and positions itself as Google’s fastest and most cost-efficient image model. It produces an image in roughly four seconds at a cost of $0.034 per 1,000 images, while retaining strong prompt adherence, character consistency, and legible in-image text rendering according to Google.
The model is live on Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform. A consumer rollout to AI Mode in Search, Google Photos, NotebookLM, and Google Ads is underway.
Product managers Alisa Fortin and Anish Nangia described the two models as complementary: developers can chain rapid image generation with Omni Flash’s video animation to build end-to-end multimedia pipelines.