bug(anthropic): ChatAnthropic sends assistant prefill for models that don't support it

### Checked other resources - [x] This is a bug, not a usage question. - [x] I added a clear and descriptive title that summarizes this issue. - [x] I used the GitHub search to find a similar question and didn't find it. - [x] I am sure that this is a bug in LangChain rather than my code. - [x] The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package). - [x] This is not related to the langchain-community package. - [x] I posted a self-contained, minimal, reproducible example. A maintainer can copy it and run it AS IS. ### Package - [x] langchain-anthropic ### Reproduction Steps / Example Code (Python) ```python from langchain_anthropic import ChatAnthropic from langchain_core.messages import HumanMessage, AIMessage llm = ChatAnthropic(model="claude-opus-4-6") response = llm.invoke([ HumanMessage(content="how are you"), AIMessage(content="I'm doing well, thanks for asking!"), ]) print(response) ``` ### Error Message and Stack Trace ``` anthropic.BadRequestError: Error code: 400 - {'type': 'error', 'error': {'type': 'invalid_request_error', 'message': 'This model does not support assistant message prefill. The conversation must end with a user message.'}, 'request_id': 'req_011CZqA2cguKkPXHv5Edm5GU'} ``` ### Description When invoking `ChatAnthropic` with `claude-opus-4-6`, passing a message list that ends with an `AIMessage` causes a 400 error from the API because the trailing AI message is sent as an assistant prefill. Older models like `claude-sonnet-4-20250514` accept prefill, but `claude-opus-4-6` does not. `ChatAnthropic` should either strip the trailing assistant message for models that don't support prefill or raise a clientside error before making the API call ### System Info ``` langchain_core: 1.0.5 langchain_anthropic: 1.0.4 anthropic: 0.73.0 Python: 3.11.13 OS: Darwin (macOS) ```