Google unveiled Nano Banana 2 Lite on June 30, 2026, presenting the model as a low-cost image generation option built for enterprise workloads. Also referred to as Gemini 3.1 Flash-Lite Image, the model is priced at $0.034 per 1,000 images and can produce a standard 1K image in 4 seconds. The company is targeting programmatic advertising, digital commerce, and automated production pipelines rather than premium creative use cases.
Built for throughput instead of top-end resolution
The model is based on the Gemini 3.1 Flash Lite architecture. To keep latency down, Google capped output at 1K resolution, leaving out the 2K and 4K options available in NB2 and NB Pro. That trade-off defines the product. Google described it as a workhorse for software engineers, ad platforms, and business systems that need large volumes of images with predictable speed and cost.
Google also highlighted improvements in several functions that matter in production settings. Nano Banana 2 Lite is designed to keep character and object appearance more consistent across image sequences, storyboards, and digital try-on tasks. It also improves text rendering inside images, which is useful for quickly generating and checking multilingual ad layouts. The company said the model can also create data visualization graphics and location-specific mockups with better accuracy.
Benchmark scores show gains over earlier models
In Google’s published benchmark results, Nano Banana 2 Lite posted a Text-to-Image Elo score of 1251. That is above the earlier NB1 score of 1151 and slightly ahead of NB Pro at 1245. For editing tasks, the model scored 1308 in single-image editing and 1294 in multi-image editing. Those numbers suggest Google is positioning the model as a practical option for broad enterprise use rather than a flagship built around maximum image fidelity.
Lowest price point in Google’s creative model lineup
Pricing is the main pressure point in this launch. Google is offering the model through a closed API approach aimed at enterprise users, which means no local deployment like some open-source alternatives, but also removes the need to manage hardware internally. Within Google’s own lineup, the price gap is clear: NB1 costs $0.039 per 1,000 images, standard NB2 costs $0.067, and NB Pro costs $0.134. Nano Banana 2 Lite comes in at $0.034, the lowest of the group.
Google’s internal estimate says the model delivers around 60% to 70% of the general capability of higher-end models. Paired with Google Workspace and the company’s broader AI product stack, that pricing strategy is aimed at locking in enterprise workflows and competing with newer rivals such as Krea 2 Turbo in automated image generation.
Available now, alongside a new video model preview
Nano Banana 2 Lite is already accessible through Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform (GEAP). Google also released a public preview of Gemini Omni Flash on the same day, a multimodal conversational model for video generation and editing. The simultaneous rollout shows Google pushing on both still-image and video tools as it expands its AI offerings for enterprise users.