Nano Banana 2 Lite

Nano Banana 2 Lite — Text-to-Image and Image Editing Model

What is Nano Banana 2 Lite?

Nano Banana 2 Lite is Google's fastest, most cost-efficient image model in the Nano Banana family, technically known as Gemini 3.1 Flash-Lite Image. It is built for rapid ideation and high-velocity developer pipelines, returning a text-to-image result in roughly four seconds. A single drop-in API handles text-to-image generation, image editing, and multi-image composition, and because it is multimodal it can return text alongside images. Despite the speed focus, it keeps reliable prompt adherence, strong character consistency, and legible in-image text rendering, making it Google's recommended upgrade from the legacy Nano Banana (Gemini 2.5 Flash Image).

Key Features

• •Fast text-to-image: outputs in about four seconds for interactive prototyping and high-volume pipelines
• •Image editing and multi-image composition: pass reference images through image_urls to edit or blend up to 14 inputs
• •1K output across 15 aspect ratios: from 1:1 feeds to 16:9 cinematic and 9:16 mobile
• •Multimodal responses: choose IMAGE only or TEXT_AND_IMAGE
• •Consistent characters across rapid generations for storyboards and try-ons
• •Legible in-image text for labels, captions, and localized ad copy
• •SynthID watermark embedded in every output

Best Use Cases

• •A/B testing ad creatives: spin up localized ad variations and layout concepts in seconds
• •Social and app content at scale: power consumer features where a user is waiting on the result
• •Storyboarding and virtual try-ons: keep character and product identity stable across frames
• •E-commerce mockups: generate product scenes and contextual layouts quickly
• •Image-to-video pipelines: produce a fast base image, then animate it with a video model
• •Rapid prototyping: explore directions cheaply before committing to a heavier model

Prompt Tips and Output Quality

Be explicit: combine subject, environment, lighting, and camera, for example "Studio photo of a red ceramic mug, softbox lighting, 50mm, shallow depth of field." Set aspect_ratio intentionally: 9:16 for stories, 16:9 for cinematic, 1:1 for feeds. Switch output_format to png for crisp edges and text. Keep thinking_level on high for complex prompts and minimal for the fastest drafts. Reuse a seed to reproduce a composition, and pass image_urls to anchor style or edit an existing image.

How to Use Nano Banana 2 Lite

Nano Banana 2 Lite (Gemini 3.1 Flash-Lite Image) is tuned for speed and volume. Use it when you need many images fast, then escalate to a heavier model only for final, high-stakes assets.

1. Craft the Prompt

• •Be explicit: combine subject, environment, lighting, and camera.

  "Studio shot of a vintage leather jacket on a wooden mannequin, softbox key light, 85mm, shallow depth of field."

• •Anchor style or edit: pass reference URLs in image_urls to guide composition or edit an existing image.
• •Repeat key details: reuse colors, props, and character names across prompts for series consistency.

2. Core Parameters

{
  "prompt": "...",
  "image_urls": [],
  "aspect_ratio": "1:1",
  "output_format": "jpg",
  "thinking_level": "high",
  "output_resolution": "1K",
  "response_modalities": "TEXT_AND_IMAGE",
  "seed": 778812
}

3. Settings by Use Case

1. •Ads and social
   • •aspect_ratio: 16:9 for banners, 9:16 for stories
   • •output_format: png for crisp text and edges
   • •thinking_level: high for layouts with copy
2. •E-commerce and try-ons
   • •aspect_ratio: 1:1 for grids, 4:5 for portrait feeds
   • •image_urls: reference catalog shots for a consistent look
3. •Storyboards and concepts
   • •aspect_ratio: 21:9 for panoramas
   • •thinking_level: minimal for the fastest iterations
   • •seed: fix a value to lock a recurring character

4. Recommended Settings

• •Fastest drafts: thinking_level minimal, output_format jpg, aspect_ratio 1:1.
• •Final polish: thinking_level high, output_format png.
• •Reproducibility: set a fixed seed and keep the prompt constant.
• •Resolution: only 1K is supported, so leave output_resolution at 1K.
• •Output type: use response_modalities IMAGE to skip the text response when you only need pixels.

5. Pro Tips

• •Batch variations: hold the prompt template constant and vary one adjective or camera angle.
• •Image-to-video: generate a base image here, then animate it with a video model.
• •Subject consistency: lock a character name and reuse it across calls.

Is Nano Banana 2 Lite open-source?

No. It is a hosted Google model (Gemini 3.1 Flash-Lite Image) accessed via API.

Can it edit images or only generate them?

Both. It supports text-to-image, image editing, and multi-image composition through image_urls.

What resolution does it output?

Outputs are generated at 1K resolution; higher 2K or 4K canvases are not supported on Lite.

How is it different from Nano Banana 2 and Pro?

Lite is speed-first; Nano Banana 2 is the generalist workhorse; Nano Banana Pro targets complex, high-fidelity work.

Are outputs watermarked?

Yes. Every image carries an invisible SynthID watermark identifying it as AI-generated.

Can I generate non-square images for ads and video?

Yes. Choose from 15 aspect ratios including 16:9, 9:16, 21:9, and ultra-wide formats.