Google released the generated image model. Nano Banana 2 LitePosition it as a light version of the HF workflow. According to the company, the new model is less delayed and a single image can be generated within about four seconds and is suitable for scenes requiring continuous modification and bulk presentation.
Lower delay and low cost.
According to Google, the price of Nano Banana 2 Lite is 0.034 per 1000 diagrams, more suitable for large-scale content drafting and rapid iterative. This version does not focus on more complex generating capabilities than in the previous version, but on speed and cost control.
The Nano Banana product line was first launched last summer, with the original version based on Gemini 3.1 Flash. The Nano Banana 2 released in February of this year emphasized greater generation capacity, including more robust image effects. Google also provides the Nano Banana Pro for higher-level use.
Replace old version of Nano Banana
Google refers to Nano Banana 2 as a generic primary power model, while Lite 2 prefers a high-volume-input task. According to the company, 2 Lite is online through Google AI Studio, Gemini API and Gemini Enterprise Argentina and will replace the original Nano Banana. The old version is now classified as “legacy model”.
In terms of product stratification, Google is splitting the same generating media product line into different slots to cover both fast bulk generation and higher-quality demand generation.
Synchronize the distribution of video models
Google announced on the same day that it would expand. Gemini Omni Flash. The model was first displayed on Google I/O this year at a cost of $0.10 per second.
In addition, Google presented a demonstration application called Omni Production Studio. According to the company, this tool allows for the conversion of static images from Omni into more “film-sense” electrician videos for such scenarios as the presentation of goods.
Google in its blog says that generating media development is often based on iterativeity. The company hopes to attract developers to complete multimedia workflows by combining video production and editing tools with rapid image generation.