> ## Documentation Index
> Fetch the complete documentation index at: https://docs.varg.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Nano Banana (Gemini)

> Google Gemini-powered image generation and editing — best default image model with reference-based editing

## Overview

Nano Banana is varg's recommended default image model, powered by Google Gemini. It excels at both text-to-image generation and reference-based image editing. Two generations are available:

| Model ID               | Generation | Mode                           | Credits | \~Cost |
| ---------------------- | ---------- | ------------------------------ | ------- | ------ |
| `nano-banana-pro`      | Gen 1      | Text-to-image                  | 5       | \$0.05 |
| `nano-banana-pro/edit` | Gen 1      | Image editing (requires input) | 5       | \$0.05 |
| `nano-banana-2`        | Gen 2      | Text-to-image + editing        | 5       | \$0.05 |
| `nano-banana-2/edit`   | Gen 2      | Image editing (explicit)       | 5       | \$0.05 |

## Quick start

<CodeGroup>
  ```typescript SDK theme={null}
  import { createVarg } from "vargai/ai"

  const varg = createVarg({ apiKey: process.env.VARG_API_KEY! })

  const result = await varg.imageModel("nano-banana-pro").generate({
    prompt: "a cozy coffee shop interior, warm lighting, plants on shelves",
    aspectRatio: "16:9",
  })

  console.log(result.image.url)
  ```

  ```bash cURL theme={null}
  curl -X POST https://api.varg.ai/v1/image \
    -H "Authorization: Bearer $VARG_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "model": "nano-banana-pro",
      "prompt": "a cozy coffee shop interior, warm lighting, plants on shelves",
      "aspect_ratio": "16:9"
    }'
  ```
</CodeGroup>

## Capabilities

### Text-to-image

Generate images from text descriptions:

```typescript theme={null}
const result = await varg.imageModel("nano-banana-pro").generate({
  prompt: "professional headshot, woman in business attire, studio lighting",
  aspectRatio: "1:1",
})
```

### Image editing

Edit an existing image using a text prompt and reference image:

<CodeGroup>
  ```typescript SDK theme={null}
  const original = Image({
    model: varg.imageModel("nano-banana-pro"),
    prompt: "woman in casual clothes standing in a park",
    aspectRatio: "9:16",
  })

  const edited = Image({
    model: varg.imageModel("nano-banana-pro/edit"),
    prompt: { text: "same woman wearing an elegant red dress", images: [original] },
    aspectRatio: "9:16",
  })
  ```

  ```bash cURL theme={null}
  curl -X POST https://api.varg.ai/v1/image \
    -H "Authorization: Bearer $VARG_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "model": "nano-banana-pro/edit",
      "prompt": "same woman wearing an elegant red dress",
      "files": [{ "url": "https://example.com/original.jpg" }]
    }'
  ```
</CodeGroup>

<Warning>
  `nano-banana-pro/edit` **requires** an input image. Plain text prompts without a reference image will fail.
</Warning>

### Multi-image reference

Provide up to 14 reference images for complex editing:

```typescript theme={null}
const edited = Image({
  model: varg.imageModel("nano-banana-pro/edit"),
  prompt: {
    text: "combine these two characters into one scene",
    images: [characterA, characterB],
  },
})
```

### Nano Banana 2 (next-gen)

The Gen 2 model supports auto aspect ratio and web search capabilities:

```bash cURL theme={null}
curl -X POST https://api.varg.ai/v1/image \
  -H "Authorization: Bearer $VARG_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nano-banana-2",
    "prompt": "a photorealistic mountain landscape at sunset"
  }'
```

## Parameters

<ResponseField name="prompt" type="string | { text: string, images: Image[] }" required>
  Text prompt for generation. For editing, provide `{ text, images }` with reference images.
</ResponseField>

<ResponseField name="aspect_ratio" type="string" default="1:1">
  Output aspect ratio. Supported: `21:9`, `16:9`, `3:2`, `4:3`, `5:4`, `1:1`, `4:5`, `3:4`, `2:3`, `9:16`. Nano Banana 2 also supports `"auto"`.
</ResponseField>

<ResponseField name="files" type="array">
  Array of `{ url: string }` objects for reference images (API). Up to 14 images for pro, up to 6 for Nano Banana 2.
</ResponseField>

<ResponseField name="provider_options" type="object">
  Supports `resolution` (`"1K"`, `"2K"`, `"4K"` — default varies by model), `safety_filter_level`, `web_search` (Nano Banana 2 only).
</ResponseField>

## Pricing

| Model                  | Credits | USD    |
| ---------------------- | ------- | ------ |
| `nano-banana-pro`      | 5       | \$0.05 |
| `nano-banana-pro/edit` | 5       | \$0.05 |
| `nano-banana-2`        | 5       | \$0.05 |
| `nano-banana-2/edit`   | 5       | \$0.05 |

The cheapest image generation with the best quality-to-cost ratio.

## Tips

* **Use Nano Banana as your default image model.** At 5 credits, it's the best value for text-to-image.
* **Image editing is the killer feature.** Generate a base image, then iteratively refine with `/edit`.
* **Perfect for video reference frames.** Generate an image with Nano Banana, then animate with Kling V3 or Seedance 2.
* **Multi-image prompts** let you combine characters, transfer styles, or composite scenes.
* **4K resolution** is available via provider options for hero images that need maximum detail.

## Related models

<CardGroup cols={2}>
  <Card title="Flux" icon="bolt" href="/models/image/flux">
    Alternative image generation. Flux Pro for highest quality.
  </Card>

  <Card title="Phota" icon="camera" href="/models/image/phota">
    Identity-preserving generation with profile-based control.
  </Card>

  <Card title="Grok Imagine Image" icon="bolt" href="/models/image/grok-imagine-image">
    Cheapest image model at 3 credits.
  </Card>

  <Card title="Seedance 2" icon="seedling" href="/models/video/seedance-2">
    Animate your Nano Banana images with Seedance image-to-video.
  </Card>
</CardGroup>