Invoking ChatAnthropic with thinking mode after tool results does not return a thinking block

[jrjenkins555]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

import asyncio

from langchain_anthropic import ChatAnthropic
from langchain_core.tools import tool


@tool
def add(num1: int, num2: int) -> int:
    """Add two numbers together."""
    return num1 + num2


convo = [
    {
        "content": "You are a helpful assistant. You help the user complete arithmetic problems. You should always respond with 'The answer is <answer>!'",
        "additional_kwargs": {},
        "response_metadata": {},
        "type": "system",
        "name": None,
        "id": None,
    },
    {
        "content": "What is 1 plus 2?",
        "additional_kwargs": {},
        "response_metadata": {},
        "type": "user",
        "name": None,
        "id": None,
    },
    {
        "content": [
            {
                "signature": "ErUBCkYIAxgCIkAvEID5bvGcsYTNq5Fzaj+TyEEyQ5nhvDPYUXpjZaBgLN/RGX+WDCh6xOEjAbMFAaVGlqs8HdD0WtvwEYS2/SPGEgxnz8HL4W++8NC7E9AaDHcpXgueNhN6W6njrSIwGRw6iPBtWzWN8Yah3b/PNiTNmEmQzlwYPKHvv8N/RV6AHMxYLLlOiW1scui5fRMrKh24n1SUtM/r/3cs7VpHFqPsMgxELP9QP9XWZZeluRgC",
                "thinking": 'The user is asking for the result of adding 1 and 2 together. I have access to an "add" function that can perform this calculation.\n\nLooking at the add function:\n- It requires two parameters: num1 and num2\n- Both parameters are integers\n- In this case, num1 = 1 and num2 = 2\n\nI\'ll use the add function to calculate 1 + 2.',
                "type": "thinking",
            },
            {
                "text": "I can help you solve this addition problem using the add function.",
                "type": "text",
            },
            {
                "id": "toolu_01KHiDW6SM1jxZ2DkcDgYrmW",
                "input": {"num1": 1, "num2": 2},
                "name": "add",
                "type": "tool_use",
            },
        ],
        "additional_kwargs": {},
        "response_metadata": {
            "id": "msg_01ADXsV6YF4FFiNNUuF1cn4Z",
            "model": "claude-3-7-sonnet-20250219",
            "stop_reason": "tool_use",
            "stop_sequence": None,
            "usage": {
                "cache_creation_input_tokens": 0,
                "cache_read_input_tokens": 0,
                "input_tokens": 446,
                "output_tokens": 188,
            },
            "model_name": "claude-3-7-sonnet-20250219",
        },
        "type": "ai",
        "name": None,
        "id": "run-83de1eef-d7f4-42f3-9d76-49a964853a54-0",
        "example": False,
        "tool_calls": [
            {
                "name": "add",
                "args": {"num1": 1, "num2": 2},
                "id": "toolu_01KHiDW6SM1jxZ2DkcDgYrmW",
                "type": "tool_call",
            }
        ],
        "invalid_tool_calls": [],
        "usage_metadata": {
            "input_tokens": 446,
            "output_tokens": 188,
            "total_tokens": 634,
            "input_token_details": {"cache_read": 0, "cache_creation": 0},
        },
    },
    {
        "content": "3",
        "additional_kwargs": {},
        "response_metadata": {},
        "type": "tool",
        "name": "add",
        "id": None,
        "tool_call_id": "toolu_01KHiDW6SM1jxZ2DkcDgYrmW",
        "artifact": None,
        "status": "success",
    },
]


async def test():
    model = ChatAnthropic(
        model="claude-3-7-sonnet-20250219",
        temperature=1,
        max_tokens=10000,
        timeout=None,
        max_retries=2,
        thinking={"type": "enabled", "budget_tokens": 5000},
    )
    print(model)
    response = await model.ainvoke(convo)
    print(response.model_dump())


if __name__ == "__main__":
    asyncio.run(test())

Description

Whenever a tool result is the most recent message, the model will not respond with a thinking block - I iteratively call an anthropic model with one tool with thinking mode enabled. I set up this script to test and used a dummy example to demonstrate this and am able to reproduce. I know thinking is enabled so I'm assuming that I should expect a thinking block in every response. I tried on latest langchain version and this did not work.

Based on docs I believe this is incorrect but I'm definitely not receiving any errors. Any pointers appreciated - thanks!

System Info

Dependencies

langchain-anthropic==0.3.13
langchain-core==0.3.59
langchain-google-genai==2.1.3
langchain-openai==0.3.14

System Information

OS: Darwin
OS Version: Darwin Kernel Version 23.4.0: Wed Feb 21 21:44:54 PST 2024; root:xnu-10063.101.15~2/RELEASE_ARM64_T6030
Python Version: 3.12.0 (main, Oct 2 2023, 20:56:14) [Clang 16.0.3 ]

Package Information

langchain_core: 0.3.59
langsmith: 0.3.32
langchain_anthropic: 0.3.13
langchain_google_genai: 2.1.3
langchain_openai: 0.3.14

Optional packages not installed

langserve

Other Dependencies

anthropic<1,>=0.51.0: Installed. No version info available.
filetype: 1.2.0
google-ai-generativelanguage: 0.6.17
httpx: 0.28.1
jsonpatch<2.0,>=1.33: Installed. No version info available.
langchain-core<1.0.0,>=0.3.53: Installed. No version info available.
langchain-core<1.0.0,>=0.3.59: Installed. No version info available.
langsmith-pyo3: Installed. No version info available.
langsmith<0.4,>=0.1.125: Installed. No version info available.
openai-agents: Installed. No version info available.
openai<2.0.0,>=1.68.2: Installed. No version info available.
opentelemetry-api: Installed. No version info available.
opentelemetry-exporter-otlp-proto-http: Installed. No version info available.
opentelemetry-sdk: Installed. No version info available.
orjson: 3.10.16
packaging: 24.2
packaging<25,>=23.2: Installed. No version info available.
pydantic: 2.11.3
pydantic<3.0.0,>=2.5.2;: Installed. No version info available.
pydantic<3.0.0,>=2.7.4: Installed. No version info available.
pydantic<3.0.0,>=2.7.4;: Installed. No version info available.
pytest: 8.3.5
PyYAML>=5.3: Installed. No version info available.
requests: 2.32.3
requests-toolbelt: 1.0.0
rich: Installed. No version info available.
tenacity!=8.4.0,<10.0.0,>=8.1.0: Installed. No version info available.
tiktoken<1,>=0.7: Installed. No version info available.
typing-extensions>=4.7: Installed. No version info available.
zstandard: 0.23.0