Google has introduced the Gemini 2.5 Flash Image model, previously known to beta users as “nanobanana.” This model is designed to provide enterprises with enhanced options for creative projects, allowing for rapid image modifications with more control compared to earlier iterations.
Gemini 2.5 Flash Image will be integrated into the Gemini app and aims to improve the consistency of character likeness across edited images. For instance, if a user uploads a photo of a pet and requests a background change or the addition of an accessory, the model will execute such requests without altering the subject’s appearance significantly.
According to Google, the latest update is intended to ensure that images of familiar people and pets maintain their likeness, as subtle changes can impact user satisfaction. Enterprises and individual users previously faced issues where minor edits could inadvertently alter the overall appearance of subjects in AI-generated images.
All images produced through Gemini will feature Google’s SynthID watermark. This model is accessible to both free and paid users of the Gemini app.
There was significant speculation about the release of a new image model, which drew attention on social media platforms. Users reported observing the model’s impressive capability to follow complex instructions, generating high-quality outputs. While early testers confirmed the model’s origins, Google had not disclosed specific plans for its future usage at that time.
The competition in the AI image editing sector is intensifying, with rivals such as Qwen and OpenAI enhancing their offerings. Additionally, Adobe has incorporated AI-powered tools in its flagship Photoshop product, intensifying the race among major players to deliver advanced image-editing capabilities.
Overall, Gemini 2.5 Flash Image represents a strategic move in the evolving landscape of multimodal AI technologies.
Source: https://venturebeat.com/ai/gemini-expands-image-editing-for-enterprises-consistency-collaboration-and-control-at-scale/