Google releases Nano Banana 2 Lite for faster, cheaper AI images

In short: Google has launched Nano Banana 2 Lite, a new AI image model that makes pictures much faster and at a lower cost than its larger options.

What happened

Google DeepMind says its new model, called Nano Banana 2 Lite, is now available across Google’s tools. It is part of the Gemini 3.1 family and its technical name is Gemini 3.1 Flash Lite Image.

The main selling point is speed. Google says it can turn a text prompt into an image in about 4 seconds in its default “low-thinking” mode. The standard Nano Banana model takes about 20 seconds for the same task, according to the report.

Google also says the model is cheaper to use through its API (an API is like a plug that lets other apps connect to the model). It averages about $0.034 per 1,000 images, and its token pricing is about half the rate of Nano Banana 2. Tokens are tiny chunks of text that get counted for billing, like charging by word count.

There are trade-offs. Google says Nano Banana 2 Lite struggles more with text in images, especially small text, and infographics are more likely to show incorrect data. It can also have trouble keeping characters consistent across multiple versions.

Google says images made with the model include SynthID watermarks, which are meant to label content as AI-made even after edits.

Why it matters

Faster and cheaper image tools can make it easier for people and businesses to try ideas quickly, like sketching rough drafts before a final design. But it may also lead to more low-quality or misleading AI images spreading online, especially when images include text or charts.

Source: Arstechnica