[Release] langchain-ai/langchain langchain==1.2.16: langchain==1.2.16

LangChain v1.2.16 release includes performance optimizations for agent state handling and streaming, new content-block-centric streaming in core, dependency updates, and documentation corrections. Key improvements focus on reducing overhead in tool dispatch, adding LRU caching for type resolution, and enhancing streaming metadata efficiency. The release also bumps related packages including Anthropic 1.4.2, OpenAI 1.2.1, and LangChain Core 1.3.0.

Key Points

• •Performance optimization: Stop inlining agent state into tool-dispatch `Send`s to reduce overhead
• •New content-block-centric streaming (v2) added to core for improved streaming capabilities
• •LRU cache implementation for type resolution to improve performance during initialization
• •Added `ls_agent_type` tag on `create_agent` calls for better agent tracking and monitoring
• •Reduced streaming metadata overhead in core for more efficient data transmission
• •Fixed Fireworks integration to properly honor `max_retries` configuration
• •Corrected import paths in agent middleware and ModelCallLimitMiddleware docstrings
• •Dependency updates: LangSmith 0.7.31, pytest 9.0.3, Pillow 12.2.0, cryptography 46.0.7
• •Added benchmark command for performance testing and monitoring
• •Coordinated releases across ecosystem: Anthropic 1.4.2, OpenAI 1.2.0/1.2.1, Core 1.3.0

Workflow Diagram

Concepts