Google DeepMind rolled out two new generative AI models aimed at developers building creative applications. Nano Banana 2 Lite targets image generation workflows where speed matters, while Gemini Omni Flash handles video creation and conversational editing. Both models are now accessible through Google AI Studio, the Gemini API, and Gemini Enterprise Agent Platform.

Nano Banana 2 Lite delivers text-to-image outputs in just four seconds, positioning it for rapid prototyping and high-volume use cases. At $0.034 per 1K-resolution image, the model offers compelling cost efficiency for developers managing operational budgets. Google positions this as the recommended upgrade path from the original Nano Banana model, citing improvements across speed, quality, and cost dimensions.

Gemini Omni Flash brings multimodal reasoning to video generation, enabling developers to create and edit videos using text, image, and video inputs. The model supports conversational editing workflows and maintains scene consistency through multimodal referencing. Priced at $0.10 per second of output—matching Veo 3.1 Fast—Omni Flash generates 10-second clips with text-action synchronization capabilities.

These releases reflect Google's push to provide specialized tools for different stages of the creative pipeline. Developers can now chain Nano Banana 2 Lite's rapid image generation with Omni Flash's video animation capabilities, creating end-to-end multimedia experiences. Demo applications in Google AI Studio illustrate practical workflows combining both models for applications like virtual travel, interior design visualization, and e-commerce video creation.