Overview
Gemini Nano Banana 2.1 is a high-efficiency model for image generation and conversational image editing. It updates Nano Banana 2 and is based on Gemini 3.6 Flash. Google announced general availability on October 6, 2026. (aistudio.google.com)
Capabilities and uses
- Accepts text, images, video and PDFs; produces images and text. The Gemini API model page specifies a 131,072-token input limit and a 32,768-token output limit. (aistudio.google.com)
- Supports image generation, multi-turn editing, Google Web and Image Search grounding, and configurable thinking levels. It supports up to 14 reference images, including references for up to four characters and ten objects. Output resolutions are 1K, 2K and 4K; 1K is the default. (aistudio.google.com)
- Suitable uses include iterative image editing, posters, infographics, diagrams and multilingual text rendering. (deepmind.google)
Limitations
The model can hallucinate, time out or render small text and long passages poorly. Character consistency, instruction following, spatial understanding, factuality and 3D reasoning are not guaranteed. The Gemini API documentation lists audio generation, code execution, function calling, structured output, URL context and context caching as unsupported. Google Cloud separately lists implicit context caching as supported on its Enterprise Agent Platform; platform capabilities should not be assumed identical. (deepmind.google)
The DeepMind model card gives a context window of up to 1 million tokens, whereas the Gemini API and Google Cloud model pages specify 131,072 input tokens. Use the limit documented for the platform being used rather than assuming the model-card figure applies to the API. (deepmind.google)
Gemini Developer API pricing
Official paid-tier prices in USD, checked November 15, 2026. The model's free-tier input and output pricing is listed as unavailable. These prices are for the Gemini Developer API, not a statement of Google Cloud platform pricing. (aistudio.google.com)
| Mode | Input, per 1 million tokens | Text and thinking output, per 1 million tokens | Image output, per 1 million tokens |
|---|---|---|---|
| Standard | $1.50 for text, image or video | $7.50 | $30.00 |
| Batch | $0.75 for text, image or video | $3.75 | $15.00 |
Google lists equivalent image-output costs per generated image of $0.0336 at 1K, $0.0504 at 2K and $0.113 at 4K for Standard; and $0.0168 at 1K, $0.0252 at 2K and $0.0567 at 4K for Batch. Input and any text or thinking output are charged separately. For Standard Google Web and Image Search grounding, the page lists 5,000 free search requests per month shared across Gemini 3 and newer models, then $14 per 1,000 requests, subject to a 1,500-request daily cap. (aistudio.google.com)