The technology giant has announced the availability of Nano Banana 2 Lite, an Artificial Intelligence (AI) model highly optimized, lightweight and fast to generate imagesand Gemini Omni Flash, a multi-modal video tool capable of generating enterprise-level productivity 10-second videos at high speed.
Both models stand out mainly for their efficiency, speed and low cost. With them, developers can create end-to-end multimedia experiences whether generating thousands of images with Nano Banana 2 Lite or editing video sequences with ‘prompts’ through Gemini Omni Flash.
Nano Banana 2 Lite (gemini-3.1-flash-lite-image) is the fastest and most powerful AI imaging model. with the best cost-efficiency ratio of the Nano Banana family, as Google has detailed in a statement on its website. These Google models were very popular among users in August of last year thanks to their ability to edit images in a very effective way.
Thus, the company recommends using the new model to replace the first version of Nano Banana (gemini-2.5-flash-image), since it has faster capabilities at the same time. generate images in less than four seconds with quality similar to Nano Banana 2which remains the most powerful and versatile model in terms of workflow and editing.
Nano Banana 2 Lite is also distinguished by its cost efficiency, as Its rate is 0.034 dollars per thousand images generatedcompared to approximately $0.067 per thousand images for the standard Nano Banana 2.
TWIN ALL FLASH
For its part, Gemini Omni Flash is presented as a multimodal video generation and editing tool capable of generating clips of up to 10 seconds from ‘prompts’, images or videos.
Presented at Google I/O, Google’s new AI model is now available for developers and stands out for its ability to perform complex edits such as from asset location to product exchange or dynamic style transfer.
Another of its virtues is that admit up to three conversational editions consecutive in the same multimedia element without losing the thread of the previous ‘prompts’. That is, it retains the format and multimedia content without introducing unwanted alterations while the user edits the final video.
For example, it is capable of replacing a Rubik’s cube with a completely transparent one in a 10-second video. without altering the rest of the scenemaintaining the fidelity of the original video.
Regarding the cost, Google charges $0.10 per second of generated videowhich equals the cost of Gemini Omni Flash with Veo 3.1 Fast (the high-speed optimized version of Google’s flagship generation model, Veo 3.1), but providing the great virtue of editing the video through natural language.
‘ANYWHERE’ DEMO TO TRY BOTH MODELS
Google has made the ‘Anywhere’ demo available to any user so they can test the capabilities of both models. After uploading a photo or taking a selfie, the demo uses the Nano Banana 2 Lite to transport the user through a series of iconic locations.
When the user clicks on an image, the Omni Flash model converts it into an animated video clip of the location that includes the user’s AI-generated photo. Other demos are available such as ‘Space Lift’, which allows you to reimagine any interior of a home by uploading an image, and ‘Omni Product Studio’, which converts static images into cinematic advertising videos.
Both models are available in public preview on Google AI Studio and the Gemini API, although the Nano Banana 2 Lite also appears on the Search AI Mode, the Gemini ‘app’, NotebookLM, Google PhotosStitch, Google Flow y Google Ads.