Positioning
Google’s efficiency-focused image generation and editing model, designed for low latency, high throughput and cost-sensitive applications. It is distinct from Nano Banana 2 and Gemini Omni Flash. (blog.google)
Capabilities and uses
- Accepts text, images, video and PDF input; produces images and text. Supports interleaved image/text generation, multi-turn image editing and thinking. (aistudio.google.com)
- Suitable for rapid visual drafting, interactive prototyping, high-volume creative pipelines, background adjustments and character-based imagery. Generated images carry SynthID watermarking and support C2PA credentials. (blog.google)
- Available through Google AI Studio, the Gemini API and Gemini Enterprise Agent Platform; Google also announced rollout across consumer products. Availability and features can differ by platform. (blog.google)
Limits
- Gemini API documentation lists a 65,536-token input limit and 4,096-token output limit. Image output supports 1K, but not 2K or 4K; 14 aspect ratios are supported. (aistudio.google.com)
- The Gemini API model documentation lists context caching, function calling, code execution, structured output, search grounding, audio generation and the Live API as unsupported. Google Cloud separately documents implicit context caching for its Agent Platform implementation, so platform-specific capabilities should not be assumed interchangeable. (aistudio.google.com)
- Fine details, small faces, spelling, factual content, translation, complex edits and character consistency can be imperfect. Review generated images and verify factual information before use. (deepmind.google)
Release date
Google’s announcement is dated June 30, 2026, and Firebase lists that date for the stable model. Google Cloud’s Agent Platform documentation separately lists June 23, 2026 as its release date. (blog.google)
Pricing
Gemini Developer API paid tier, in USD per 1 million tokens; checked November 16, 2026. These are not necessarily Agent Platform prices. (aistudio.google.com)
| Mode | Input: text/image/video | Output: text and thinking | Output: image |
|---|---|---|---|
| Standard | $0.25 | $1.50 | $30.00 |
| Batch | $0.125 | $0.75 | $15.00 |
A 1K image uses 1,120 output tokens, equivalent to $0.0336 per image under standard pricing or $0.0168 per image under batch pricing. Input and other applicable output tokens are charged separately. The pricing page lists the free tier as unavailable for this model. (aistudio.google.com)