Gemini Nano Banana improves image editing consistency and control at scale for enterprises – but is not perfect

Google has unveiled its latest innovation, the Gemini 2.5 Flash Image, a model previously known as 'nanobanana' during its beta phase. This new tool brings enhanced options for enterprises looking to elevate their creative projects, allowing for quick alterations to images while maintaining greater control than earlier versions. Integrated into the Gemini app, this model builds upon its predecessor by enhancing the native image editing capabilities. It excels at preserving the likeness of subjects across various images, ensuring a consistent portrayal. For instance, if a user uploads a picture of their pet and requests a background change or the addition of a hat, Gemini 2.5 Flash Image will execute these edits without distorting the pet's features. In a blog post by David Sharon, the lead for multimodal generation at Gemini Apps, and Nicole Brichtova, the product lead for Gemini image, Google emphasized the importance of subtlety in image editing. They stated, "When editing pictures of yourself or loved ones, even minor flaws can feel off. This update aims to keep your friends, family, and pets looking true to form." Despite its advancements, some users have expressed concerns about the model's performance when making adjustments to AI-generated images. For example, while it follows commands to reposition a subject, it may inadvertently alter facial features. All images processed through Gemini will feature Google’s Synth ID watermark, ensuring transparency. The introduction of the model sparked considerable buzz on social media, with users on LM Arena discussing its capabilities, including following intricate, multi-step instructions with notable precision. Early testers confirmed suspicions that the 'nanobanana' model originated from Google, although the company had yet to officially announce its intentions at that time. As excitement builds, Google faces stiff competition from other AI image editing providers, including Qwen and OpenAI, which recently integrated AI image editing into ChatGPT. Adobe, a long-standing leader in image editing, has also added its Firefly model to its suite of applications. The emergence of Gemini 2.5 Flash Image represents a significant step in the ongoing battle among tech giants to produce more sophisticated and realistic image editing tools. This innovative model not only allows users to modify images but also facilitates multi-turn editing and the blending of styles between different pictures. By incorporating such features into the chat platform, enterprises can streamline their workflow, making it easier to enhance images or graphics without juggling multiple windows.