Google launches Nano Banana 2 Lite for fast AI images and Gemini Omni Flash for video via API

Google has added two generative media models to its lineup, both oriented toward developers who need fast, cost-effective outputs via API. Nano Banana 2 Lite handles image generation, and Gemini Omni Flash extends the Gemini family into video - covering both generation and editing through text prompts.

Nano Banana 2 Lite is positioned as a speed-first image model, completing generations in approximately four seconds and priced at $0.034 per image. That combination of low latency and relatively modest per-unit cost makes it a practical option for applications that need to produce images at scale or in near-real-time, rather than waiting on heavier, more expensive models.

Gemini Omni Flash is the more significant structural addition. Video generation has been available through Google's consumer-facing tools, but bringing it to the API opens the capability to developers building their own products and workflows. The model supports both creating video from a text prompt and editing existing footage through natural language instructions - a pairing that broadens its utility beyond simple clip generation.

Google is actively promoting a chained workflow that combines the two releases: use Nano Banana 2 Lite to quickly produce a source image, then pass that image to Gemini Omni Flash to animate or extend it into video. This kind of model chaining is a common pattern in generative media pipelines, and framing it explicitly suggests Google sees the two as complementary parts of a single production path rather than separate standalone tools.