Google is quietly testing a new image-generation model called Nano Banana 2 Flash, representing a shift toward making advanced AI more accessible for everyday use. Built for speed, efficiency, and lower costs, early signs suggest it could become the fastest image AI the company has developed so far.
Nano Banana 2 Flash is part of Gemini’s Flash line-up, which focuses on low-latency generative AI. While Google has not officially announced the model, reports from late 2025 and early 2026 indicate it is already being tested internally. Nano Banana 2 Flash points to the company’s next step in making AI tools quicker, cheaper, and more widely accessible.
While the model won’t be as powerful as Nano Banana Pro (Gemini 3 Pro Image) — Google’s top-end image generation and editing model — it is expected to be significantly faster and cheaper to run, making it ideal for quick tasks, lightweight apps, and rapid creative workflows.
The model was spotted on X by MarsForTech, a known source for Gemini leaks, and later appeared in internal code. Reports suggest the model may carry the codename “Mayo,” continuing Google’s condiment-themed naming scheme.
Although Google hasn’t officially announced the model, its repeated appearances in internal code and user reports suggest a launch could be approaching. Some users have even claimed brief access, a pattern that often signals Google is testing readiness for a wider rollout.
If launched, Nano Banana 2 Flash would give users a faster, more accessible way to generate images, without replacing Google’s more powerful premium models. Instead, it would offer a clear choice between speed and precision — and could soon power everyday image creation for developers and even free-tier users.
