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

# Gemini 4 Hydrogen

> Gemini 4 Hydrogen for fast coding and multimodal agent workflows, with a shared 1M context window and four thinking levels through MixRoute.

Gemini 4 Hydrogen is Google's speed- and efficiency-focused Gemini 4 model for everyday coding, interactive agents and multimodal work. It accepts text, images, video, audio and PDFs, and produces text.

## Model specifications

| Field | Description |
| - | - |
| `model` | gemini-4-hydrogen |
| `Input` | Text, images, video, audio, PDF |
| `Output` | Text |
| `Context window` | 1,048,576 tokens |
| `Thinking levels` | low / medium / high / xhigh |

The provider's 1,048,576-token window is shared by input, thinking and output. A standard request has a 65,536-token combined thinking/output cap. Longer generation uses a separate continuation workflow; these examples do not establish MixRoute support for long decode or Interactions.

## Request examples

Set `MIXROUTE_API_KEY` to your MixRoute API key.

### Gemini native API

```bash theme={null}
curl --request POST "https://api.mixroute.ai/v1beta/models/gemini-4-hydrogen:generateContent" \
  --header "Authorization: Bearer $MIXROUTE_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
  "contents": [
    {
      "role": "user",
      "parts": [
        {
          "text": "Reply with exactly OK."
        }
      ]
    }
  ],
  "generationConfig": {
    "maxOutputTokens": 256,
    "thinkingConfig": {
      "thinkingLevel": "low"
    }
  }
}'
```

### OpenAI-compatible Chat Completions

```bash theme={null}
curl --request POST "https://api.mixroute.ai/v1/chat/completions" \
  --header "Authorization: Bearer $MIXROUTE_API_KEY" \
  --header "Content-Type: application/json" \
  --data '{
  "model": "gemini-4-hydrogen",
  "messages": [
    {
      "role": "user",
      "content": "Reply with exactly OK."
    }
  ],
  "max_tokens": 256,
  "reasoning_effort": "low"
}'
```

## Request parameters

| Field | Description |
| - | - |
| `contents` | Conversation content with role and parts; text is placed in parts\[].text. |
| `generationConfig.maxOutputTokens` | Combined budget for reasoning and output text, not a target answer length. |
| `generationConfig.thinkingConfig.thinkingLevel` | low, medium, high (provider default), or xhigh. Thinking cannot be disabled with minimal. Prefer `thinkingLevel` over the deprecated `thinkingBudget`, and leave enough output budget for both reasoning and the final answer. |
| `model / messages / max_tokens / reasoning_effort` | Use gemini-4-hydrogen with messages. The example requests low reasoning effort and a small output budget. |

## Integration notes

* The examples use text input. Multimodal formats, function calling and structured output follow their respective Gemini contracts; route limits still apply.
* Omit `temperature`, `topP`, `topK` and `candidateCount`: Google ignores custom sampling values for this model. Explicit context caching through `cachedContent` is not supported.
* The presence of built-in Google tools in the provider model specification does not establish that every hosted tool, Files API operation or continuation flow is available through MixRoute.

Availability and pricing depend on your MixRoute route. See the [Model Marketplace](https://console.mixroute.ai/models).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.