Google DeepMind has opened developer access to Nano Banana 2 Lite and Gemini Omni Flash. The release gives builders new tools for fast image generation and multimodal video-oriented workflows.

The announcement matters because generative media is becoming a standard part of AI application stacks rather than a separate creative tool. Lower-latency and API-accessible models can make image and video features easier to embed in products.

For developers, the practical test will be whether the new models combine speed, controllability and cost well enough for production use.