### Check for existing issues

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

### What happened?

## Summary

When calling `sap/gemini-3.5-flash` through LiteLLM, the request appears to complete successfully, but LiteLLM fails while transforming the SAP response into a `ModelResponse`.

The SAP response contains `reasoning_content` as a list of structured reasoning blocks, while LiteLLM's `ModelResponse.reasoning_content` field expects a string.

This results in a Pydantic validation error before the completion response is returned to the caller.

## Environment

* LiteLLM version: `1.92.0` * Python version: `3.12` * Provider: SAP AI Core / SAP Generative AI Hub * Model: `sap/gemini-3.5-flash` * Pydantic version: `2.13.x`

## Minimal reproduction

```python from litellm import completion

response = completion( model="sap/gemini-3.5-flash", messages=[ { "role": "user", "content": "Hi!" } ] )

print(response) ```

## Expected behavior

LiteLLM should return a valid `ModelResponse`, for example:

```python response.choices[0].message.content ```

If SAP returns structured reasoning blocks, L…