TECHNOLOGY

LangChain Core 1.4.0: Deprecation Fix and Dependency Updates

LangChain Core 1.4.0: Deprecation Fix and Dependency Updates

LangChain Core has bumped to version 1.4.0, and while it's a minor release, it packs a meaningful fix for developers working with deprecation warnings. The update addresses an issue where the @deprecated decorator was triggering an eager import of pydantic.v1, potentially slowing down startup times or causing unnecessary dependencies to load. Alongside this, several core dependencies—urllib3, mistune, jupyter-server—have been updated to their latest versions. Here’s what changed and why it matters.

What Changed

The headline fix is in libs/core: a commit titled fix(core): avoid eager pydantic.v1 import in @deprecated. Previously, using the decorator would force Python to import pydantic.v1 even if you weren't using it. That’s now deferred. The team also set the deprecation since parameter to 1.3.3 to match the actual release version—a small but important consistency tweak. On the dependency front, you get urllib3 2.7.0, mistune 3.2.1, and jupyter-server 2.18.0. The merge from v1.4 into master suggests this release is stable and ready for production.

Why It Matters

This fix isn't flashy, but it’s the kind of under-the-hood polish that keeps LangChain reliable. For teams running large deployments or using the deprecated decorator extensively, avoiding an eager import can shave off milliseconds from cold starts—and every little bit counts in serverless or edge environments. The dependency bumps also patch security issues and improve performance, though the changelog doesn’t specify CVEs. Personally, I’m glad to see the team paying attention to versioning consistency; nothing’s worse than a deprecation warning pointing to a non-existent release. If you're on an older version, this is a low-risk upgrade that won't break your code but will keep things tidy.

Official Source: https://github.com/langchain-ai/langchain/releases/tag/langchain-core%3D%3D1.4.0