Gemini 3.1 Flash Image Preview
Neural Network
Gemini 3.1 Flash Image Preview is a Google model for generating and editing images from text descriptions and references.
Specifications
- Input
- text, images
- Output
- images
- Resolution
- 1K, 2K, 4K
- Aspect ratio
- 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 16:9, 9:16, 21:9
- Capabilities
- Batch API
Image prompt
(per 1K tokens)
How it works Gemini 3.1 Flash Image Preview?
Gemini 3.1 Flash Image Preview, also known as Nano Banana 2, is a generative Google model that creates and edits images: it combines Pro-level visual quality with Flash-series speed. It works with both text descriptions and uploaded references, making it suitable for precise edits to existing frames and maintaining character consistency across image series. In BotHub, the model is accessible from Russia without a VPN or foreign card; payment is made with Russian cards in rubles based on actual token usage, and unused balances do not expire. Over 250 neural networks are available in one window, allowing you to compare results with other models in a few clicks. A unified OpenAI-compatible API lets you integrate image generation into your product and switch models without rewriting integrations. Designers create concepts and mood boards, marketers prepare creatives for ads and social media, online stores create product cards and visuals, and development teams embed image generation into their own services via API.Frequently asked questions about Gemini 3.1 Flash Image Preview
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
nano-banana-2-preview from google-gemini generates images from text descriptions and processes uploaded images based on references. It is suitable for visuals for posts and ads, product cards, concepts, and illustrations, while BATCH mode helps prepare series of frames in bulk.
The model outputs images created from text descriptions or based on an uploaded frame. Specific resolutions and file formats are not recorded in our data, so please refer to the results in the interface. Context is 65,536 tokens, output limit is 32,000 tokens per request.
There is no specific note about step-by-step reasoning in our data, and we will not deny such a possibility. The model's task is images: the prompt is processed by the language part, so long descriptions with scenes, styles, and details are parsed coherently. If you need detailed reasoning, use a text model nearby.
In our data, function calling and structured JSON output for nano-banana-2-preview are not noted — this is not a denial, but an absence of a label. If your pipeline logic requires tools and strict schemas, build it on a text model: BotHub's unified OpenAI-compatible API allows switching without rewriting integrations.
Images — yes: input of images is confirmed, image-to-image modes and reference-based editing work. There are no notes about audio and video in our data, so you shouldn't count on them, but we cannot claim otherwise. Text plus image is the main scenario.
We do not publish data on prompt languages for this model, so we won't promise anything specific — check with your own request, it takes a couple of minutes. On the BotHub side, everything is in Russian: interface, support, payment with Russian cards in rubles, and access from the Russian Federation without a VPN.