AI
Google's new Nano Banana 2 Lite image model is its fastest and cheapest yet
Google DeepMind has released Nano Banana 2 Lite (technically named Gemini 3.1 Flash Lite Image), a fast and low-cost image-generation model.
Key takeaways
- Nano Banana 2 Lite offers the best balance of quality and speed.
- The model can generate an image from text in about 4 seconds in default low-thinking mode.
- API rates for the new model are half the rate of Nano Banana 2.
- Users rate Nano Banana 2 Lite outputs almost as highly as the non-Lite versions on Arena.ai.
- The model has more trouble with text, infographics, and character consistency across iterations.
Google DeepMind has released Nano Banana 2 Lite (technically named Gemini 3.1 Flash Lite Image), a fast and low-cost image-generation model. It can generate images from text in about 4 seconds in its default low-thinking mode, compared to 20 seconds for the standard Nano Banana. The model is available across the Google ecosystem, Google AI Studio, and the Gemini API, costing developers an average of $0.034 per 1K images ($0.25 per 1M input tokens and $1.50 per 1M output tokens). Additionally, Google is expanding access to Gemini Omni Flash, its video-generation model, to the Gemini API and Google AI Studio.
By the numbers
- 4s
- Time to generate an image in default low-thinking mode
- 20s
- Time to generate an image in standard Nano Banana
- $0.034
- Average cost per 1K images generated
- $0.25
- API rate per 1M input tokens
- $1.50
- API rate per 1M output tokens
How it unfolded
- Google announced Gemini Omni Flash at I/O
- Nano Banana 2 Lite released across Google ecosystem
Turn stories like this into views
Ravenclip finds the AI news, makes the video, and posts it before attention moves on.
Common questions
- What happened with Gemini 3.1 Flash Lite Image?
- Nano Banana 2 Lite offers the best balance of quality and speed.
- Where can I read the original report?
- Read the full report at ars_technica.
