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

# Gemini native format

## Gemini content generation (native format)

Call Gemini multimodal models (e.g. Nano Banana) with native `generateContent`: text, images, and optional “thinking” traces.

***

### Basics

* **Endpoint**: `POST /v1beta/models/{model}:generateContent/`
* **Path**: `model` — e.g. `gemini-3-pro-image-preview`.
* **Auth**: `Bearer Token`
* **Content-Type**: `application/json`

***

### Request body

| Field                     | Type      | Required | Description                                |
| :------------------------ | :-------- | :------- | :----------------------------------------- |
| **contents**              | `array`   | **Yes**  | Conversation turns.                        |
| └─ **role**               | `string`  | No       | e.g. `user`.                               |
| └─ **parts**              | `array`   | **Yes**  | Message parts.                             |
| └─ └─ **text**            | `string`  | **Yes**  | Prompt / text input.                       |
| **generationConfig**      | `object`  | **Yes**  | Generation settings.                       |
| └─ **responseModalities** | `array`   | **Yes**  | `TEXT`, `IMAGE`, etc.                      |
| └─ **thinkingConfig**     | `object`  | No       | Thinking / chain-of-thought.               |
| └─ └─ **includeThoughts** | `boolean` | No       | Return model “thoughts” before the answer. |
| └─ **imageConfig**        | `object`  | **Yes**  | Image settings.                            |
| └─ └─ **aspectRatio**     | `string`  | **Yes**  | e.g. `16:9`, `1:1`, `4:3`.                 |
| └─ └─ **imageSize**       | `string`  | **Yes**  | e.g. `4K`, `1024x1024`.                    |

***

### Response

**200 OK**

| Field                | Type     | Description                              |
| :------------------- | :------- | :--------------------------------------- |
| **candidates**       | `array`  | Model outputs.                           |
| └─ **content**       | `object` | Role + `parts` (text and/or image).      |
| └─ **finishReason**  | `string` | e.g. `STOP`.                             |
| └─ **safetyRatings** | `array`  | Safety scores.                           |
| **usageMetadata**    | `object` | Token usage (prompt, candidates, total). |

***

### Example (image)

```json theme={null}
{
  "contents": [
    {
      "role": "user",
      "parts": [
        {
          "text": "draw a futuristic city at sunset"
        }
      ]
    }
  ],
  "generationConfig": {
    "responseModalities": ["TEXT", "IMAGE"],
    "imageConfig": {
      "aspectRatio": "16:9",
      "imageSize": "4K"
    },
    "thinkingConfig": {
      "includeThoughts": true
    }
  }
}
```

***

### Tips

1. **Thoughts**: `includeThoughts: true` surfaces how the model interprets the prompt—useful for debugging complex prompts.
2. **Multimodal**: With both `TEXT` and `IMAGE`, you often get explanatory text then the image.
3. **Model id**: Ensure `{model}` matches what your key can access.