### Check for existing issues

- [x] I have searched the existing issues and checked that my issue is not a duplicate.

### What happened?

LiteLlm treats Cerebras models as though they are OpenAI compatible, but they're only *mostly* compatible. Specifically, OpenAI supports both `tools` and `response_schema` arguments in the same call, while Cerebras doesn't. This is leading to errors where LiteLlm (in my case, via ADK) is trying to call the model with both and getting an exception.

### Steps to Reproduce

1. Make a completion call like the following:

``` acompletion( model="cerebras/gpt-oss-120b", messages=[{"role": "system", "content": "You are a helpful agent"}, {"role": "user", "content": [{"type": "text", "text": "hello"}] ], tools=[...], # a non-empty list of tools response_format={"type": "json_schema", {...}}, # some json schema here api_base="https://api.cerebras.ai/v1", api_key="...cerebras api key...", } ```

2. Receive error response

### Relevant log output

```shell litellm.exceptions.BadRequestError: litellm.BadRequestError: CerebrasException - "tools" is incompatible with "response_format" ```

### What part of LiteLLM is this about?

S…