A new standardized tool-calling interface for AIMessage has been introduced to facilitate seamless interaction with external data sources through large language models (LLMs), allowing developers to create sophisticated applications that manipulate resources like databases, files, and APIs. This interface, designed to be fully backward compatible with models having native tool-calling support, includes methods such as ChatModel.bind_tools(), AIMessage.tool_calls, and create_tool_calling_agent(), providing a consistent approach across different LLM providers like OpenAI, Anthropic, and Gemini. These methods allow users to define tool availability, access tool invocations, and build agents with any tool-calling model. The introduction of this standardized interface is expected to enhance the flexibility and usability of LLMs by enabling easier switching between providers and facilitating agent construction within the LangChain framework, while the recently released ChatModel.with_structured_output() interface complements this by offering structured outputs for information extraction tasks. Users are encouraged to update their langchain_core and partner package versions to utilize these new features, which aim to streamline workflows and improve integration capabilities.