Gemini 3 Pro Image (Nano Banana Pro): Google's Image Generation and Editing Model, Features and Pricing

What Gemini 3 Pro Image is

Gemini 3 Pro Image, which Google also calls Nano Banana Pro, is an image generation and editing model from Google DeepMind, introduced on 20 November 2025. It is built on Gemini 3 Pro and uses Gemini's reasoning and real-world knowledge to visualise information. Google's documentation describes it as best for complex graphic design, high-fidelity product mockups and factual data visualisations that need accurate text rendering and grounding in Google Search. It launched as gemini-3-pro-image-preview and later became generally available as gemini-3-pro-image. It is a proprietary model with no published weights. On LMArena it has no LMArena rating.

What Google highlights

Google says Nano Banana Pro renders accurate, legible text directly in images, from a short tagline to a long paragraph, in multiple languages and with a wider variety of textures, fonts and calligraphy, and that it can translate and localise text inside an image. It can connect to Google Search to visualise real-time information such as weather or sports, and create infographics and diagrams from supplied content or real-world facts. It blends up to 14 input images while keeping the resemblance of up to 5 people. Creative controls include localised editing, camera angle, focus, colour grading and day-to-night lighting changes, with a range of aspect ratios and 2K and 4K output.

Specifications

On the Gemini API, gemini-3-pro-image takes image and text input and returns image and text output, with an input limit of 65,536 tokens and an output limit of 32,768 tokens. Google lists support for thinking and grounding with Google Search, and the model can be used through the Batch API as well as flex and priority inference. Google's model page gives November 2025 as its latest update and links a DeepMind model card.

Availability and watermarking

At launch Nano Banana Pro rolled out in the Gemini app, where free users received limited quotas before reverting to the original Nano Banana and Google AI Plus, Pro and Ultra subscribers got higher quotas. It also came to AI Mode in Search for Pro and Ultra subscribers in the US, NotebookLM, Google Ads, Google Slides and Vids for Workspace customers, Flow for Ultra subscribers, the Gemini API, Google AI Studio, Google Antigravity and Vertex AI. All its images carry Google's imperceptible SynthID watermark. Images from free and Google AI Pro users also carry a visible Gemini sparkle, which Google removes for Ultra subscribers and in Google AI Studio.

Pricing

On the Gemini API's paid tier, image output is priced at $120 per million tokens. An output image from 1K up to 2K resolution uses 1,120 tokens, about $0.134 per image, and a 4K image uses 2,000 tokens, about $0.24 per image. Text and image input costs $2.00 per million tokens, and an input image counts as 560 tokens, about $0.0011. The model is not available on the free tier. Grounding with Google Search includes 5,000 free search requests a month shared across Gemini 3 and newer models, then costs $14 per 1,000 requests.

Official Resources