LangChain shipped langchain-core 1.6.2 on September 4, 2026, following 1.6.1. The release adds async tool support for OpenAI and patches mutation bugs in Google GenAI and Bedrock Converse standard content handling.
LangChain released langchain-core 1.6.0, adding standard model exception types and fixing several issues including postponed annotations in StructuredTool and serialization of RunnablePick. The update improves tool schema resolution and Windows portability.
LangChain released langchain-core 1.6.0, adding standard model exception types and fixing several issues including postponed annotations in StructuredTool and serialization of RunnablePick. The update improves tool schema resolution and Windows portability.
LangChain shipped langchain-openai 1.6.1 on GitHub, fixing routing of gpt-5.6-sol to the responses API and adding support for async tools, configuration_update, and Azure AD auth with OpenAI 3.8. The release matters to developers building on OpenAI models through LangChain, since it also corrects reasoning_effort_levels metadata for gpt-5 and gpt-5.1.
LangChain shipped langchain-openai 1.6.1 on GitHub, fixing routing of gpt-5.6-sol to the responses API and adding support for async tools, configuration_update, and Azure AD auth with OpenAI 3.8. The release matters to developers building on OpenAI models through LangChain, since it also corrects reasoning_effort_levels metadata for gpt-5 and gpt-5.1.
LangChain released langchain-openai 1.5.0, adding support for the OpenAI 3.0 SDK and bumping the langgraph floor in lockfiles. The update is a minor release following 1.4.3.
LangChain released langchain-openai 1.5.0, adding support for the OpenAI 3.0 SDK and bumping the langgraph floor in lockfiles. The update is a minor release following 1.4.3.
LangChain released langchain-core v1.5.0, introducing `reasoning_effort` as a standard chat model parameter. This allows developers to control the amount of reasoning a model applies, enabling cost-performance trade-offs. The release also includes dependency bumps for soupsieve and mistune.
LangChain released langchain-core v1.5.0, introducing `reasoning_effort` as a standard chat model parameter. This allows developers to control the amount of reasoning a model applies, enabling cost-performance trade-offs. The release also includes dependency bumps for soupsieve and mistune.
LangChain published langchain-openai 1.7.2 on GitHub, a patch release following 1.7.1. The update fixes handling of OpenAI Responses API streams so items stay on one block when output_index drifts.
LangChain published langchain-openai 1.7.2 on GitHub, a patch release following 1.7.1. The update fixes handling of OpenAI Responses API streams so items stay on one block when output_index drifts.