Prime Highlight
- Google has launched Gemini 2.5 Flash Image, offering faster, more consistent, and higher-quality image generation and editing.
- The model introduces character consistency and simple text-based editing, enabling seamless creative workflows for both professionals and casual users.
Key Facts
- Gemini 2.5 Flash Image supports multi-image fusion, background edits, object removal, and pose adjustments through plain text commands.
- Now available in preview on Gemini API, Google AI Studio, and Vertex AI, with costs set at $30 per million tokens and $0.039 per image.
Background
Google has released Gemini 2.5 Flash Image, the latest version of its image generation and editing tool. The new model, nicknamed nano-banana, brings better editing options, stronger consistency, and more accurate results compared to earlier versions.
Gemini 2.5 Flash Image is an upgrade version of Gemini 2.0 Flash and is part of a series named Gemini 2.5, which was fast, but the quality of the picture was limited. The new model has been created to be beneficial in both a fast test and a professional creative work.
The character consistency is one of the main additions as it helps the AI retain the same subject or object as it appears in various edits or prompts. It can be helpful to make pictures of products or create brand designs, or to make characters move around.
Simple editing using text commands is also supported by the model. Users can ask it to change backgrounds, remove objects, adjust poses, or edit small details using plain language. A new multi-image fusion feature lets users combine parts of different images into one.
Gemini 2.5 Flash Image also uses world knowledge, which allows it to understand drawings, adapt templates, and help in education tasks where text and images are linked.
Industrial designer Thomas Broen, who tested the tool, said it was quick and reliable, even able to restore the original image after edits.
Gemini 2.5 Flash Image is now in preview on Gemini API, Google AI Studio, and Vertex AI, and will be stable soon. The cost is approximately $30 per 1 million tokens of output and about $0.039 per image generated. Google AI Studio has also launched new templates that could give developers and creators a simpler method to test.
