Google has added two generative AI models aimed at developers who need faster image creation and API-based video tools. Nano Banana 2 Lite focuses on low-cost, high-speed image generation, while Gemini Omni Flash brings text-prompt video generation and editing into the Gemini API and Google AI Studio.
The bigger story is not just that Google has another image model and another video model. It is that the company is encouraging developers to use them together: generate a reference image first, then turn that image into a short clip.
What Nano Banana 2 Lite changes for image generation
Nano Banana 2 Lite is built around speed and throughput. Google says it can generate an image in four seconds, with pricing of $0.034 per image at 1K resolution. In the API, the model is listed as gemini-3.1-flash-lite-image.
That makes the model especially relevant for teams that need to test many visual ideas quickly. A developer building a product workflow, design assistant, ad tool, or creative prototype can use a faster model to explore options before deciding which outputs deserve more refinement.
Google says the model is designed for fast ideation and high-throughput developer pipelines. Even with that speed focus, the company says Nano Banana 2 Lite supports reliable prompt following, consistent character rendering, and readable text in generated images.
The model is also moving beyond developer tools. Google says Nano Banana 2 Lite is rolling out across AI Mode in Google Search, the Gemini app, NotebookLM, Google Photos, Stitch, Google Flow, and Google Ads.
How it fits into the Nano Banana family
Nano Banana 2 Lite expands the Nano Banana family to three production models. The practical difference is positioning: developers can choose based on speed, image quality, cost, or control.
- Nano Banana 2 Lite is the fast, low-cost option for rapid image generation.
- Nano Banana 2 (Gemini 3.1 Flash Image) is positioned as the all-rounder with the best balance of quality and cost.
- Nano Banana Pro (Gemini 3(.1) Pro Image) is aimed at complex, professional work, with what Google calls the strongest control and most advanced reasoning.
Google considers the original Nano Banana (Gemini 2.5 Flash Image) outdated. For teams using Google image models in production, the new lineup creates a clearer decision path: use the lighter model when speed matters most, use the balanced model when output quality and cost both matter, and use the Pro model when a more demanding visual task needs stronger control.
Gemini Omni Flash brings video into the API
Gemini Omni Flash was first shown at Google I/O and is now available through the Gemini API and Google AI Studio. It combines Gemini's multimodal reasoning with video generation and editing, giving developers a way to create and revise clips through text prompts.
The model is priced at $0.10 per second of video output, matching Veo 3.1 Fast. It can generate and edit videos up to ten seconds long through the API.
Google highlights three strengths for Gemini Omni Flash: conversational video editing through natural language, support for mixed inputs such as text, images, and video, and use of Gemini's world knowledge during generation. The model can also sync text and graphics directly with actions in a video.
Those capabilities point to a different kind of video workflow from traditional editing. Instead of adjusting every element manually, a developer can build an interface where a user describes the change they want, provides reference material, and receives a generated or edited clip.
The limits developers need to know
Gemini Omni Flash is not presented as a complete replacement for every video generation workflow. The current version only generates ten-second clips. Audio references and scene extensions are not yet supported in the API.
There are also limits around video references. The API schema accepts video references up to three seconds long, but Google says the model does not process them correctly yet. Character consistency across scene changes or camera movements is still limited too.
Those constraints matter for production planning. Short clips, product motion, concept previews, and simple animated references fit the current shape of the model better than long-form scenes or complex continuity-heavy video.
Why Google wants the models chained together
Google sees the largest benefit in combining Nano Banana 2 Lite with Gemini Omni Flash. The workflow is straightforward: generate images quickly with Nano Banana 2 Lite, then pass those images into Gemini Omni Flash as references for animation.
The Interactions API, which is now Google's default AI API, preserves session history and context. Google says it allows up to three consecutive edits, which gives developers a way to build iterative creative tools rather than one-shot generators.
Google has provided three demo apps to show the combined workflow. "Anywhere" places users at famous landmarks via selfie and animates the result. "Space Lift" creates interior design concepts from room photos and turns them into video. "Omni Product Studio" converts static product images into e-commerce videos.
Both models use SynthID watermarks to tag AI-generated content, according to Google. Verification is available through the Gemini app, Gemini in Chrome, or Google Search. Nano Banana 2 Lite and Gemini Omni Flash are available now in Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform.