Progress often arrives like a banana peel underfoot – sudden, disruptive, and impossible to ignore.
This is evident in the visual arts, where initial scepticism often gives way to innovation. Painters once dismissed photography, claiming it lacked the skill of traditional art. Photographers, in turn, were wary of Adobe Photoshop’s ability to manipulate reality through pixels. This cycle of doubt and adaptation is familiar to agency veterans. Now, the relentless tide of innovation has reached digital artists. Rather than a threat, tools like Nano Banana represent an evolution – not replacing creativity, but expanding its possibilities for greater efficiency, imagination, and opportunity.
Mystery Banana
AI has dominated industry headlines in recent years. Each new text-to-image release comes with glossy demos, press kits, and orchestrated hype campaigns. For those following AI developments closely, the industry is saturated with press releases, curated demos, and paid social media posts. Amid this AI frenzy, something unusual occurred in mid-August. A new image model quietly appeared on platforms like LMArena, where users anonymously compare AI models side by side. There was no announcement or press release, just raw performance on display.
The model’s quality was undeniable.
Initially lacking a name or backstory, the model’s jaw-dropping results spoke for themselves. This mystery became its genius. The AI community promoted it, relentlessly testing, analysing outputs, and speculating about its origins. Before marketers could shape the narrative, the product was validated in the toughest testing ground possible – public opinion.
When revealed, the tool was a Google project, officially named Gemini 2.5 Flash Image. However, it was dubbed Nano Banana due to two quirks: ‘Nano’ for its compact 500 MB size, optimised for CPUs and mobile chips, and ‘Banana’ from a humorous glitch that added bananas to unrelated images. Rather than a bug, this became its brand, transforming into a meme and a story. The playful nickname made cutting-edge AI feel approachable, breaking down barriers to entry in a way no technical product label could.
Not so Nano
While MidJourney sets the standard for artistic image generation, Nano Banana excels in a different, arguably more practical dimension: image editing.
For advertisers and image professionals, this is where the model shines. Whether shifting a scene, swapping a product, or adjusting a viewing angle, Nano Banana performs with remarkable precision. It seamlessly adds people to compositions, generates convincing product and fashion mock-ups, transfers styles from one visual world to another, restores damaged images, extends canvases, and edits text embedded in visuals – a longstanding challenge for traditional AI tools. Such tasks would once require painstaking hours in Photoshop, if possible at all, and still might not achieve the same quality. Nano Banana’s ability to follow prompts faithfully and maintain character consistency means fewer iterations before achieving a usable result.
Paint by Numbers
All this comes at a startlingly accessible price: roughly $30 USD per million output tokens, or just under four cents per image. In an industry where each round of revisions adds time and cost, Nano Banana makes high-level image editing not only faster but also more affordable.
More than just a tool, Nano Banana signals a shift in creative approaches within advertising. By dramatically reducing the time, cost, and challenges of image production, it liberates creative teams to focus on ideas rather than limitations. Budgetary constraints, once limiting great ideas, are now less of a concern, allowing imaginations to soar.
For advertising agencies embracing this technology, the future promises coexistence of speed, quality, and innovation and the only boundaries are the limits of their own creativity. In a world where every pixel matters, Nano Banana empowers creatives to dream bigger, faster, and bring visions to life like never before.