# Using Google Gemini

4AICode's Gemini models use the OpenAI-compatible `chat/completions` protocol. Users only need to use their own `sk-4ai-...` token and change `model` to `google/gemini-*`. The platform routes requests through AI Studio or Vertex AI, depending on availability.

## Available Models

| Model | Recommended Use Cases |
|---|---|
| `google/gemini-3.1-pro-preview` | Complex reasoning, long context, preview features |
| `google/gemini-3-flash-preview` | Faster responses, preview features |
| `google/gemini-2.5-flash` | Everyday Q&A, lightweight coding, low cost |
| `google/gemini-2.5-pro` | Complex tasks, code analysis, consistent quality |


## Model Routing

Use the `google/gemini-*` model name in your request. The platform manages upstream routing; current model availability and prices are shown in the console. Upstream quota limits or service interruptions can affect availability.

## Python

```python
from openai import OpenAI

client = OpenAI(
    base_url="https://relay.4aicode.com/v1",
    api_key="sk-4ai-xxxxxxxxxxxxxxxx",
)

resp = client.chat.completions.create(
    model="google/gemini-2.5-flash",
    messages=[
        {"role": "user", "content": "Introduce Gemini in one sentence"},
    ],
)

print(resp.choices[0].message.content)
```

## TypeScript / Node.js

```ts
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://relay.4aicode.com/v1",
  apiKey: "sk-4ai-xxxxxxxxxxxxxxxx",
});

const resp = await client.chat.completions.create({
  model: "google/gemini-2.5-flash",
  messages: [{ role: "user", content: "Write a TypeScript hello world" }],
});

console.log(resp.choices[0].message.content);
```

## curl

```bash
curl https://relay.4aicode.com/v1/chat/completions \
  -H "Authorization: Bearer sk-4ai-xxxxxxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-2.5-flash",
    "messages": [
      {"role": "user", "content": "ping"}
    ]
  }'
```

## Streaming Output

```bash
curl https://relay.4aicode.com/v1/chat/completions \
  -H "Authorization: Bearer sk-4ai-xxxxxxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-2.5-flash",
    "stream": true,
    "messages": [
      {"role": "user", "content": "Write a Go hello world"}
    ]
  }'
```

The response format remains OpenAI SSE:

```text
data: {"choices":[{"delta":{"content":"Hello"}}]}
data: [DONE]
```

## Image Input

Gemini currently supports base64 data URL image input:

```python
resp = client.chat.completions.create(
    model="google/gemini-2.5-flash",
    messages=[{
        "role": "user",
        "content": [
            {"type": "text", "text": "Describe this image"},
            {
                "type": "image_url",
                "image_url": {
                    "url": "data:image/png;base64,iVBORw0KGgo..."
                },
            },
        ],
    }],
)
```

Remote `https://...` image URLs currently return a clear error. Download the image and convert it to a base64 data URL before sending it.

## Common Errors

| Error | Resolution |
|---|---|
| `model not supported` | Confirm that the model name is one of those listed above |
| `no available google account` | The upstream Gemini service is temporarily unavailable; contact support |
| `quota exceeded` | Insufficient balance or quota; recharge or contact your team administrator |
| `upstream error` | Google upstream instability or API key quota issues; try again later |