Nano Banana

nano banana is an experiment that demonstrated Google's Gemini 1.5 model's ability to correctly identify the content of an image with just 48 bytes. The feat showed how generative artificial intelligence is becoming more efficient, able to work with minimal amounts of data and still deliver accurate answers.
This evolution marks an important leap for anyone building digital products with low-code, no-code, or embedded AI, enabling faster, lighter, and more scalable experiences.

With the trend toward lower data usage and improved inference, technologies like this become essential for building AI solutions even in contexts with limited infrastructure or incomplete data flows.

‍

‍

Currently, Gemini 1.5 is tied to the Google Cloud platform and accessed via API. The cost varies according to the volume of calls and processing, making it ideal for companies looking to scale AI solutions securely and in a customizable way.

Although still in early adoption across much of the market, access to models like Gemini is expected to become more affordable in the coming months.

‍

Gemini 1.5 Starter Plan
Price: US$0.000125/token
Limit: Billed per token generated and interpreted
Hosting: Exclusively via Google Cloud
Extra features: Use of image, text, and multimodal data with limited input

Gemini 1.5 Business Plan
Price: Variable by volume
Limit: Broader tokenization and support for larger contexts
Hosting: Google Cloud + hybrid environments
Extra features: Optimized performance, reduced latency, and enterprise SLAs

Gemini 1.5 Enterprise Plan
Price: On request
Limit: Expansive tokenization, advanced multimodal support
Hosting: Self-hosted (in beta) or via Google Cloud
Extra features: Custom models, advanced security, Vertex AI integration, dedicated support

‍

Google's nano banana is not just a curious technical demo but a landmark that signals the future of artificial intelligence: lighter, more efficient, and more accessible.
With the advance of models like Gemini 1.5, a new possibility opens up for companies looking to apply generative AI in no-code and low-code flows, without depending on large volumes of data or complex structures.

For anyone building digital products, now is the time to experiment, prototype, and explore new AI-assisted experiences, even in contexts with little information.
At Flowcode, we keep a close eye on these movements to turn technological advances into practical solutions. If you want to understand how to take advantage of this new standard of AI efficiency, talk to us.

‍