Gemini 3.1 Flash Lite Image (Nano Banana 2 Lite): Google's Low-Latency Image Model, Limits and Pricing

What Gemini 3.1 Flash Lite Image is

Gemini 3.1 Flash Lite Image, also called Nano Banana 2 Lite, is an image generation and editing model from Google, released as a stable model on 30 June 2026 under the name gemini-3.1-flash-lite-image. Google describes it as the efficiency specialist of its image generation family, built for ultra-low latency and low-cost generation and editing. It targets end-to-end latency under two seconds and significantly reduced TPU compute costs, which Google says suits high-volume developer use and real-time consumer apps. On LMArena's text-to-image leaderboard Gemini 3.1 Flash Lite Image has no LMArena rating.

Capabilities

The model takes text, images, video and PDFs as input and returns images and text, with native support for interleaved generation and editing: text in, text and images out, or an image plus text in, text and images out. Google highlights fast multi-turn local edits such as swapping colours, making stickers and adjusting backgrounds across several turns of one conversation, and says it keeps the character consistency of the original Nano Banana. It supports thinking at the minimal and high levels, and the Batch API, Flex and Priority inference. Function calling, Google Search and Maps grounding, code execution and context caching are not supported.

Sizes and limits

Gemini 3.1 Flash Lite Image is optimised for lower resolutions and generates only at 1K, 1024 by 1024 pixels; 2K and 4K are not supported. It supports a set of 14 aspect ratios, including 1:1, 3:2, 2:3, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9 and 21:9. Input is limited to 65,536 tokens and output to 4,096 tokens. Every image carries an always-on SynthID watermark and C2PA content credentials.

Pricing

There is no free tier for this model in the Gemini API. On the paid tier, input costs $0.25 per million tokens for text, images or video. Output costs $1.50 per million tokens for text and thinking and $30 per million image tokens, which Google says equals $0.0336 per 1K image. Batch pricing halves these figures, to $0.125 input, $0.75 text output and $15 per million image tokens, or $0.0168 per 1K image. At the standard rate 1,000 images cost about $33.60, and about $16.80 through Batch. Text and thinking output are billed at the text rate, so a response that mixes an image with text is charged at both rates. Google states that paid-tier data is not used to improve its products.

Where it fits

Google's image family also includes Nano Banana 2 and Nano Banana Pro, which support higher resolutions. Gemini 3.1 Flash Lite Image can be tried in Google AI Studio and called through the Gemini API, and Google DeepMind publishes a model card for it. Developers call it with the model code gemini-3.1-flash-lite-image. Google lists its latest update as June 2026, and the model page marks it as a stable release rather than a preview.

Official Resources