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

# Google Gemini

> Use Gemini models through the Nolma gateway

# Google Gemini via Nolma

## Setup

Gemini requests go through the `/gemini` prefix:

```python theme={null}
import openai  # Gemini supports OpenAI-compatible API

client = openai.OpenAI(
    api_key=os.environ["GEMINI_API_KEY"],
    base_url="https://gateway.nolma.ai/gemini",
    default_headers={
        "NM-Key": "nm_live_abc123",
        "NM-Agent": "my-agent",
    }
)
```

## Supported models

| Model            | Tier    | Cost (input / output per 1M tokens) |
| ---------------- | ------- | ----------------------------------- |
| gemini-2.5-pro   | Premium | $1.25 / $10.00                      |
| gemini-2.0-flash | Economy | $0.10 / $0.40                       |

## Native Gemini SDK

You can also use the native Gemini SDK with a custom endpoint:

```python theme={null}
import google.generativeai as genai

# Configure to use Nolma gateway
genai.configure(
    api_key=os.environ["GEMINI_API_KEY"],
    transport="rest",
    client_options={
        "api_endpoint": "https://gateway.nolma.ai/gemini"
    }
)
```

<Note>
  When using the native Gemini SDK, pass Nolma headers via the request metadata or use the OpenAI-compatible interface for simpler integration.
</Note>