Google adds Nano Banana 2 Lite and Gemini Omni Flash for developers

Google has expanded its generative-media lineup with two developer-focused models, pushing further into cheap, high-volume image and video creation. Both went live on 30 June 2026 across Google AI Studio, the Gemini API and the Gemini Enterprise Agent Platform.

The first, Nano Banana 2 Lite, is billed as Google’s fastest and most cost-efficient image model, carrying the codename gemini-3.1-flash-lite-image. It produces text-to-image outputs in about four seconds and is priced at $0.034 per 1,000 images at 1K resolution, a rate aimed at teams generating images at scale. Google highlights reliable prompt adherence, consistent character rendering and legible in-image text as selling points. Beyond developer channels, the model also feeds consumer surfaces including AI Mode in Search, the Gemini app, NotebookLM, Google Photos and Google Ads.

The second, Gemini Omni Flash, arrives in public preview under the identifier gemini-omni-flash-preview. It handles high-quality video generation together with conversational editing, letting users refine clips through back-and-forth instructions rather than re-rendering from scratch. It accepts text, image and video inputs and supports what Google calls text and action synchronisation. For now it generates clips up to 10 seconds long, with longer durations promised later, and is priced at $0.10 per second of video output.

The pairing underlines how aggressively the major labs are competing on price-performance for creative tooling, not just reasoning benchmarks. By separating a lightweight image workhorse from a preview-stage video model and exposing both through the same API surface, Google is courting developers building media pipelines who need predictable per-asset costs.

For regional agencies, e-commerce teams and app builders, the low per-image and per-second pricing lowers the barrier to embedding on-demand visuals and short video into products, though the 10-second cap and preview status of Omni Flash mean production use will still require testing against specific workloads.

Sources: Google

Drafted with AI assistance and reviewed by the IT Magazine news desk. Facts verified against the linked sources. Spotted an error? See our Corrections Policy.