Nano Banana
In plain wordsThe nickname, later Google's official branding, for its Gemini image models, a viral hit in 2025 for editing photos from simple instructions while keeping people looking like themselves.
The popular name for Google DeepMind's image generation and editing models built into its Gemini multimodal modelsmultimodal large language model, n.: An AI language model that can also take in — and sometimes produce — pictures, sound or video, so you can show it a photo or speak to it, not just type.: Gemini 2.5 Flash Image (August 2025), Gemini 3 Pro Image, branded Nano Banana Pro (November 2025), and Gemini 3.1 Flash Image, branded Nano Banana 2 (February 2026); as of October 2026 the newest was Nano Banana 2.1. Unlike stand-alone text-to-image modelstext-to-image model, n.: An AI system that creates a picture from a written description, such as 'a fox reading a newspaper, in watercolor', usually within seconds., they create and edit images within a conversation, drawing on the language model's world knowledge, and are noted for instruction-based image editinginstruction-based image editing, n.: Changing a picture by telling an AI what to change — 'make it night', 'remove the car', 'give her a red coat' — instead of using editing tools by hand. that keeps a person's or object's identity consistent across edits, for blending several reference images and, in later versions, for legible text. Outputs carry SynthIDSynthID, prop. n.: Google's system for hiding an invisible watermark in pictures, video, audio and text made by its AI tools, so they can later be recognized as AI-generated. watermarks. Within weeks of launch, Google reported more than 10 million new Gemini app users and over 200 million image edits, helped by a viral trend of turning selfies into photorealistic toy-figurine images. In August 2026 Google retired its ImagenImagen, prop. n.: Google's family of AI image generators, first shown in 2022, whose research showed that better understanding of language yields pictures that match descriptions more faithfully. models from the Gemini API in favor of these models.
Nano Banana swapped the jacket and the background but kept her face exactly the same.
Origin
'Nano Banana' was the codename under which the first model was tested anonymously on the LMArena leaderboard from 12 August 2025. Google confirmed its identity at public release on 26 August 2025 and then adopted the nickname as branding; it reportedly derives from nicknames of Google DeepMind product manager Naina Raisinghani.
Related
- ImagenImagen, prop. n.: Google's family of AI image generators, first shown in 2022, whose research showed that better understanding of language yields pictures that match descriptions more faithfully.
- instruction-based image editinginstruction-based image editing, n.: Changing a picture by telling an AI what to change — 'make it night', 'remove the car', 'give her a red coat' — instead of using editing tools by hand.
- text-to-image modeltext-to-image model, n.: An AI system that creates a picture from a written description, such as 'a fox reading a newspaper, in watercolor', usually within seconds.
- multimodal large language modelmultimodal large language model, n.: An AI language model that can also take in — and sometimes produce — pictures, sound or video, so you can show it a photo or speak to it, not just type.
- SynthIDSynthID, prop. n.: Google's system for hiding an invisible watermark in pictures, video, audio and text made by its AI tools, so they can later be recognized as AI-generated.
- LMArenaLMArena, prop. n.: A public website that ranks AI chatbots by blind taste test: visitors put one question to two unnamed models, pick the reply they prefer, and millions of such votes produce the rankings.
Sources
- Google (2026). Migrate from Imagen to a Gemini Image model ('Nano Banana'). Firebase AI Logic documentation. firebase.google.com/docs/ai-logic/imagen-models-migration
- Google (2026). Gemini models. Gemini API documentation. ai.google.dev/gemini-api/docs/models
Mentioned in
ImagenImagen, prop. n.: Google's family of AI image generators, first shown in 2022, whose research showed that better understanding of language yields pictures that match descriptions more faithfully., instruction-based image editinginstruction-based image editing, n.: Changing a picture by telling an AI what to change — 'make it night', 'remove the car', 'give her a red coat' — instead of using editing tools by hand., VeoVeo, prop. n.: Google's video-generating AI, which since 2025 creates short clips complete with matching sound and dialogue from a written prompt.