Google Opens Generative Media APIs: Nano Banana 2 Lite at $0.034/K Images, Gemini Omni Flash at $0.10/sec Video

TL;DR: Google released two generative media APIs on June 30 — Nano Banana 2 Lite (fastest/cheapest image model at $0.034/K images) and Gemini Omni Flash (video generation at $0.10/sec) — both callable from the Gemini API now.

Google shipped two generative media APIs on June 30, both available via Google AI Studio and the Gemini API:

Nano Banana 2 Lite replaces the original Nano Banana as the recommended image generation model. It generates images in ~4 seconds at $0.034 per 1,000 images — competitive with the cheapest third-party image APIs while staying in the Google ecosystem. Available across AI Studio, Gemini API, and the Gemini Enterprise Agent Platform.

Gemini Omni Flash brings video generation and conversational editing to the API. Pass text, images, or video clips; get back short video clips. Clips are capped at 10 seconds at launch, priced at $0.10/sec of output (matching Veo 3.1 Fast). Use case: programmatic video generation pipelines, video-to-video editing, and multimodal agent workflows that need to produce video artifacts.

Why it matters: Both models are API-first — not limited to Google's consumer apps. If you're building multimodal pipelines (image gen as a step, video summarization, AI-generated product media), Nano Banana 2 Lite's pricing makes image generation a cheap add-on step rather than a budget line item.

Action required: The gemini-2.5-flash-preview-09-2025 model is deprecated July 9. Migrate to gemini-2.5-flash now if you're using the preview endpoint.

Sources: Google blog: Start building with Nano Banana 2 Lite and Gemini Omni Flash · Google Cloud Blog · Gemini API changelog