← All models

Gemini 3 Pro Image

Published Publisher Googlegemini-3-pro-image

Overview

Gemini 3 Pro Image, also called Nano Banana Pro, is Google’s image generation and editing model built on Gemini 3 Pro. It is designed for complex, high-fidelity image creation and multi-turn editing. (docs.cloud.google.com)

Capabilities and uses

  • Accepts text and image inputs and produces text and images. It supports reasoning, system instructions and implicit context caching. (docs.cloud.google.com)
  • Supports image generation and conversational editing, including text rendering in images and combining reference images. Example uses include marketing assets, product imagery, virtual try-ons, infographics and media production. (blog.google)
  • Supports up to 14 input images. The documented limits are 65,536 input tokens and 32,768 output tokens. (docs.cloud.google.com)

Limitations

  • Audio and video inputs are not supported. Video input mentioned in the announcement applies to Nano Banana 2, not this model. (docs.cloud.google.com)
  • At the May 28, 2026 announcement, 1K and 2K output were generally available; 4K output remained in preview. Generated images contain a SynthID watermark. (cloud.google.com)
  • The requested number of output images is not guaranteed. Structured output is not supported in the Google Cloud model documentation. (ai.google.dev)

Release

Google first introduced Nano Banana Pro on November 20, 2025. The gemini-3-pro-image release became generally available on May 28, 2026 through Gemini Enterprise Agent Platform. (blog.google)

Pricing

Currency: USD. Pricing checked November 16, 2026. The following rates are for the Gemini Developer API paid tier, not a promise of identical pricing on other platforms or regional endpoints. (ai.google.dev)

Service Input Output
Standard $2.00 per 1 million text or image tokens; approximately $0.0011 per input image $12.00 per 1 million text and thinking tokens; $120.00 per 1 million image tokens, approximately $0.134 per 1K or 2K image and $0.24 per 4K image
Batch $1.00 per 1 million text tokens; approximately $0.0006 per input image $6.00 per 1 million text and thinking tokens; $0.067 per 1K or 2K image and $0.12 per 4K image

The listed image equivalents use 560 tokens per input image, 1,120 output tokens per 1K or 2K image, and 2,000 output tokens per 4K image. The model has no Gemini Developer API free tier. Google Search grounding may incur separate charges. (ai.google.dev)

Official sources