### Check for existing issues

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

### What happened?

I can't figure out how to setup bedrock-mantle models in the proxy while I'm using an IAM Role/Policy to give permissions to the EC2 instance running the litellm proxy.

### Steps to Reproduce

If I define the model as such:

- model_name: gemma-4-31b litellm_params: model: bedrock_mantle/google.gemma-4-31b aws_region_name: us-east-1

I get the following:

Error fetching response:Error: 500 litellm.APIConnectionError: Bedrock_mantleException - {"error":{"code":"invalid_api_key","message":"Missing 'authorization' or 'x-api-key' header","param":null,"type":"permission_denied_error"}}. Received Model Group=gemma-4-31b Available Model Group Fallbacks=None

GoogleAI suggested I add a bogus api_key and it would fail-over to the role/policy...

- model_name: gemma-4-31b litellm_params: model: bedrock_mantle/google.gemma-4-31b aws_region_name: us-east-1 api_key: "uses-iam-role"

but that didn't work as it seems to just fail because it's a bad key:

Error fetching response:Error: 500 litellm.APIConnectionError: …