Google has simply launched Gemini 2.5 Flash Picture, a brand new picture modifying and era instrument woven into the Gemini app. Often known as the cryptic “nano banana” throughout testing, this mannequin has already impressed early customers, topping leaderboards and setting the bar excessive for what AI can do with visuals.
For months, whispers of “nano-banana” have been circulating on-line and on testing platforms like LMArena, the place it appeared anonymously and wowed customers. Google has now confirmed these rumors, and the thriller mannequin is certainly Gemini 2.5 Flash Picture, constructed by the DeepMind crew. In contrast to earlier AI picture editors that usually struggled to maintain a topic’s id intact throughout modifications, this mannequin is all about consistency. Add a photograph of your pal, your canine or your self and you may put them in a Nineteen Sixties diner or a futuristic cityscape and the topic will nonetheless be themselves.
Sale
Google Pixel 10 – Obsidian – 128 GB with $100 Amazon Reward Card
- Order the brand new Google Pixel 10 and get an Amazon Reward Card. Legitimate 8/20/2025 till 9/4/2025 at 11:59pm PT, whereas provides final and topic to…
- Return of Pixel 10 with out present card leads to a cost. Provide can’t be mixed, is non-transferable and never legitimate for money or money equal
- If a qualifying merchandise in your order is returned, you may be reimbursed for what you come back, minus the worth of the present card
Enhancing with Gemini 2.5 Flash Picture is so simple as typing a sentence. Have to swap out a plain T-shirt for a Hawaiian one? Simple. The mannequin understands pure language prompts so that you don’t must be a photograph modifying professional to get professional outcomes. It may even add coloration to black and white photos or change an individual’s pose with a couple of phrases.
Past single picture changes, the mannequin is nice at combining many photos into one scene. Add a photograph of your self and one other of your pet after which ask Gemini to make a portrait of you each stargazing on a mountain. The outcomes are surprisingly reasonable with the mannequin retaining face options and fur patterns all through the mix.
Multi-turn modifying is one other cool function which lets you edit a picture step-by-step. Begin with an empty room, then paint the partitions blue, add a bookshelf and eventually a armchair – all with easy textual content directions. Every change builds on the earlier one whereas holding the general scene intact. This appears like working with an artist who understands your concept and retains it irrespective of what number of modifications you make.
Google’s mannequin doesn’t simply cease at modifying current pictures; it might probably apply types to photographs. Use the feel of flower petals to rebuild a pair of rainboots, or a butterfly’s wing sample to create a garment. This appears like a magic wand for creation, permitting you to play with aesthetics in ways in which have been beforehand unique to designers.
The Gemini 2.5 Flash Picture mannequin is offered to each free and paid customers of the Gemini app, with free customers capable of create as much as 100 edits every day and paid customers getting ten occasions that. Builders can entry it via the Gemini API, Google AI Studio and Vertex AI, for $0.039 per picture generated. Adobe has already jumped on board, integrating the mannequin into Firefly and Categorical, with full entry for Categorical customers beginning September 1.
Each picture created or edited with this mannequin has a visual “AI” watermark and an invisible SynthID digital watermark so it’s AI. That is about accountable use, addressing the issues round misuse in an period the place reasonable fakes are a rising drawback.
