LangChain releases v1.3.13 and langchain-openai v1.3.5 with explicit prompt caching
LangChain released versions 1.3.13 and langchain-openai 1.3.5, adding explicit prompt caching support for OpenAI models. This allows developers to mark reusable prompt prefixes for reduced latency and cost. The update also includes a new 'meta' extra for langchain-meta integration and refreshed model profiles.
Entities
Related
LangChain Fireworks partner package v1.4.4 released with cached prompt token reporting
LangChain released version 1.4.4 of the langchain-fireworks partner package on July 9, 2026. The update adds reporting of cached prompt token usage and includes dependency bumps and documentation clarifications.
Developer seeks prompt caching optimization for GitHub Copilot Agent Mode with GPT-5.6
A developer is trying to optimize prompts for GitHub Copilot Agent Mode, noting that GPT-5.6 models make prompt caching more valuable due to specific Cache Read and Write costs. They reference OpenAI's API documentation for prompt caching (1024-token prefix, 128-token increments, identical prefix matching, short-lived in-memory caches, optional 24-hour extended caches) but cannot find whether GitHub Copilot exposes the same behavior or has its own orchestration layer.
LangChain releases v1.3.12 with interrupt propagation and shell middleware fix
LangChain released version 1.3.12, fixing interrupt propagation through ToolRetryMiddleware and avoiding shared process-group kill in shell middleware. The companion langchain-openai v1.3.4 suppresses Pydantic serializer warnings on structured output parsed fields.
LangChain-Mistralai v1.1.6 adds citation metadata support
LangChain released version 1.1.6 of its langchain-mistralai integration, adding support for surfacing citation metadata from Mistral AI chat responses. The update also includes refreshed model profiles and dependency bumps for vcrpy and langsmith.
LangChain-OpenRouter releases v0.2.6 with custom HTTP header support
LangChain-OpenRouter v0.2.6 is released on 2026-07-05, adding support for `default_headers` for custom HTTP header injection and refreshing model profile data. The update is a minor version bump from 0.2.5.