Google's 'Nano Banana' AI Image Model: The Quirky Origin Story

Google's AI development often involves internal codenames that are far from conventional, and the story behind the 'Nano Banana' moniker for its Gemini image generation model is no exception. This seemingly random name has a surprisingly personal origin, adding a touch of whimsy to the cutting-edge technology. Let's delve into the story of how Nano Banana came to be and its connection to Google's broader AI efforts.

The Genesis of 'Nano Banana'

The name 'Nano Banana' wasn't initially intended for public consumption. It began as a temporary codename for Arena.ai, a platform where AI models are tested and refined. This highlights the iterative and experimental nature of AI development, where even the naming conventions can be fluid and informal.

A Product Manager's Inspiration

The creative force behind the 'Nano Banana' name is Naina Raisinghani, a Product Manager at Google. Drawing inspiration from her own nicknames, she coined the term, injecting a bit of personality into the project. This human element in the development process serves as a reminder that even the most advanced AI technologies are shaped by individual creativity and preferences.

From Arena.ai to Gemini

The Nano Banana model was first uploaded to Arena.ai on August 12, 2025, marking a significant milestone in its development. After a period of testing and refinement, it was officially launched on August 26, 2025. While internally known as Nano Banana, Google's technical documentation refers to it as Gemini 3 Pro Image Preview. This is important context when reviewing Google's official documentation.

The Broader AI Landscape

Google isn't the only tech giant using playful codenames for its AI projects. Meta, for example, is reportedly working on an AI model nicknamed 'Avocado.' This trend suggests a lighthearted approach to internal project naming within the competitive AI landscape. The use of codenames allows teams to experiment and iterate without the pressure of a formal product name early in development.

Why This Matters

Understanding the origins and naming conventions of AI models provides valuable insight into the development process. It humanizes the technology and reveals the often-unseen creative decisions that shape these powerful tools. Furthermore, knowing the internal codenames can be helpful when tracking the progress of specific AI projects within companies like Google. For example, if you are using Google Gemini, understanding its evolution from projects like Nano Banana can provide valuable context.

Sources

• SlashGear: Why Google's Nano Banana Pro Image Model Has Such A Weird Name