[BUG] Using gemini-2.0-flash-thinking-exp model, the Cat outputs only the thinking steps, omits the actual answer.

[amagnolo]

Describe the bug
When using the gemini-2.0-flash-thinking-exp model, the Cat outputs only the thinking steps, not the actual answer.

To Reproduce

1. select gemini-2.0-flash-thinking-exp as LLM model
2. ask any question

Expected behavior
Show the model thinking steps, then the actual answer to the user question.

Additional context
The gemini 2.0 non chain-of-thought variant (gemini-2.0-flash-exp) works correctly with cheshire cat.

[pieroit]

We're not suppporting reasoning models yet, thanks for opening the issue