Implementation:Langchain ai Langchain BaseChatModel Invoke With Tools

Knowledge Sources	LangChain
Domains	NLP, Tool_Use
Last Updated	2026-02-11 00:00 GMT

Overview

Concrete tool for invoking chat models with tool schemas and receiving structured tool call responses provided by langchain-core.

Description

The BaseChatModel.invoke() method follows the same invocation pipeline (input preparation, caching, rate limiting, _generate) but when tools are bound, the _generate() method includes tool schemas in the API request. The provider's response is parsed into an AIMessage with a tool_calls attribute containing a list of ToolCall objects (each with name, args, and id fields).

Usage

Call invoke() on a model that has had bind_tools() applied. Check response.tool_calls to determine if the model wants to call tools.

Code Reference

Source Location

• Repository: langchain
• File: libs/core/langchain_core/language_models/chat_models.py (invoke); libs/partners/openai/langchain_openai/chat_models/base.py (_generate with tool processing)
• Lines: chat_models.py L389-437 (invoke); base.py L1379-1478 (_generate)

Signature

def invoke(
    self,
    input: LanguageModelInput,
    config: RunnableConfig | None = None,
    *,
    stop: list[str] | None = None,
    **kwargs: Any,
) -> AIMessage:

Import

from langchain_openai import ChatOpenAI
from langchain_core.messages import AIMessage

I/O Contract

Inputs

Name	Type	Required	Description
input	LanguageModelInput	Yes	User messages (tools are already bound via bind_tools)
stop	list[str] or None	No	Stop sequences

Outputs

Name	Type	Description
return	AIMessage	Message with tool_calls list containing ToolCall objects with name, args, and id

Usage Examples

Checking for Tool Calls

from langchain_openai import ChatOpenAI
from langchain_core.tools import tool

@tool
def get_weather(city: str) -> str:
    """Get weather for a city."""
    return f"Sunny in {city}"

llm = ChatOpenAI(model="gpt-4o-mini")
llm_with_tools = llm.bind_tools([get_weather])

response = llm_with_tools.invoke("What's the weather in London?")

if response.tool_calls:
    for tc in response.tool_calls:
        print(f"Tool: {tc['name']}, Args: {tc['args']}, ID: {tc['id']}")
        # Tool: get_weather, Args: {'city': 'London'}, ID: call_abc123
else:
    print(response.content)

Related Pages

Implements Principle

• Principle:Langchain_ai_Langchain_Tool_Invocation_With_Tools