# Switch AI Models Without Rewriting Your OpenAI SDK Integration

> Source: <https://dev.to/routerbase_com/switch-ai-models-without-rewriting-your-openai-sdk-integration-28gb>
> Published: 2026-07-16 02:17:58+00:00

If your application already uses the OpenAI Python SDK, you can keep the same client and point it at RouterBase by changing the base URL.

This gives the application one request interface while model selection stays configurable.

`openai`

packageInstall the SDK:

```
pip install --upgrade openai
```

Set the key in your shell:

```
export ROUTERBASE_API_KEY="sk-rb-..."
```

Do not place an API key in browser code, mobile binaries, or a public repository.

``` python
import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["ROUTERBASE_API_KEY"],
    base_url="https://routerbase.com/v1",
)
```

The important line is `base_url`

. The SDK remains the same; requests go through the RouterBase OpenAI-compatible endpoint.

Choose a model ID from the current RouterBase model catalog and keep it in configuration instead of hard-coding it throughout the application.

``` python
import os

model_id = os.environ.get(
    "ROUTERBASE_MODEL",
    "google/gemini-2.5-flash",
)

response = client.chat.completions.create(
    model=model_id,
    messages=[
        {
            "role": "user",
            "content": "Explain idempotency in one paragraph.",
        }
    ],
)

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

To test a different compatible chat model, change `ROUTERBASE_MODEL`

and run the same request again.

A successful API response is not enough to justify moving production traffic. Test candidate models against the same tasks and rubric.

```
candidates = [
    "google/gemini-2.5-flash",
    # Add other current RouterBase chat model IDs here.
]

prompt = "Return a JSON object with keys: summary and risk."

for candidate in candidates:
    result = client.chat.completions.create(
        model=candidate,
        messages=[{"role": "user", "content": prompt}],
    )
    print(candidate, result.choices[0].message.content)
```

In a real evaluation, record correctness, latency, retry rate, and the cost of a successful task.

An OpenAI-compatible endpoint does not mean every model has identical behavior.

Before switching models, verify:

Keep the integration stable, but keep the evaluation model-specific.

[RouterBase](https://routerbase.com) provides one OpenAI-compatible API for GPT, Claude, Gemini, and 200+ AI models. Check the live catalog before using a model ID in production.