Google has launched a brand new image-focused mannequin referred to as Gemini 2.5 Flash Picture, designed to provide and refine visuals by easy textual content instructions.
The replace was formally launched on August 26 and is already stay throughout all Gemini interfaces.
With this rollout, customers can carry out detailed edits, reminiscent of adjusting an individual’s stance or combining completely different visuals, with out altering key particulars like facial options or background components.
Do you know?
Subscribe – We publish new crypto explainer movies each week!
NEAR Protocol Defined: Newbie’s Information to NEAR (Animated)
Earlier than its official reveal, the software had surfaced on LMArena, a website the place the general public can check experimental synthetic intelligence (AI) fashions. Recognized by that title underneath the alias “nano-banana”, the mannequin gained consideration for its clean and dependable enhancing.
One of many standout options is its skill to maintain a constant look for a topic throughout a number of pictures. This may be helpful in fields like promoting, the place a product or character wants to look the identical in several scenes.
The mannequin also can pull from a number of reference visuals and perceive diagrams, which permits it to create correct visuals primarily based on a mixture of sources.
To assist forestall misuse, all pictures generated by the mannequin carry an invisible SynthID signature, together with metadata that identifies them as AI-generated.
Gemini 2.5 Flash Picture is offered not solely inside Google’s platform but in addition by OpenRouter and fal.ai. It is usually built-in into Google Cloud’s service providing, with a price of $30 for each million output tokens, which interprets to round $0.04 per picture.
On August 13, Google rolled out an replace to Gemini AI. What’s new on this model? Learn the total story.
