Google’s new Nano Banana 2 Lite image model is its fastest and cheapest yet

There are plenty of AI image-generation models out there these days, but the ones capable of producing quality output tend to be slow and expensive. Google DeepMind says its new image model, known as Nano Banana 2 Lite, offers the best balance of quality and speed. It is available throughout the Google ecosystem today, allowing images to be created in less time than Google’s improved models.

The new model is part of the Gemini 3.1 family—it’s technically called the Gemini 3.1 Flash Lite Image. On the one hand, Google says this model is ideal for exploring ideas and “rapid-fire” prototyping, applications in which quality may lag. However, the company has also provided some examples aimed at showing how close the Nano Banana 2 Lite can get to the quality of its other imaging models.

In addition to the examples, Google also has Elo scores available from Arena.ai, which shows that users rate the Nano Banana 2 Lite output about the same as the non-Lite versions. However, vibemarking doesn’t always focus on details that can make AI images look silly upon closer inspection. Google notes that the Nano Banana 2 Lite has more trouble with text, especially if it’s very small, and infographics are more likely to contain incorrect data. Characters and people may also show poor consistency across iterations.

but this is Fast. You can go from text to image in about 4 seconds in the default low-thinking mode. The standard Nano Banana takes about 20 seconds to generate the same images. Speed ​​and efficiency mean that developers accessing Nano Banana 2 Lite through the API will have to pay very little. Google says the average is $0.034 per 1K image. API rates are $0.25 for 1M input tokens and $1.50 for 1M output tokens. This is half rate for the Nano Banana 2. For Nano Banana Pro, the input token is just over $2 for 1M, but the output pricing is $12 (eight times more).