# What changed in @langchain/groq

> 15 changes in @langchain/groq 0.3 to 1 from its official release notes, upgrade guides and advisories: 3 breaking, 3 security, 5 new APIs and options. Each one cited. Sources checked 2026-10-19. HTML: https://unstale.sh/npm/@langchain/groq/

To see which of these apply to a repository, run `npx @unstale/cli init` in it.

## @langchain/groq 1

### Breaking changes

- The peer dependency now requires @langchain/core >=1.1.30, preventing module-resolution failures caused by incompatible older versions. (1.3.1) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.3.1)
- Tightens the peer dependency range to ^1.1.30 because ChatGroq imports export subpaths introduced in that version; earlier versions can cause module-not-found build failures. (1.3.0) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.3.0)
- When aborted during streaming with streaming callback handlers, invoke() throws ModelAbortError with accumulated partialOutput. (1.0.4) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.0.4)

### Security

- LangChain: MongoDBChatMessageHistory query injection can allow cross-session access (1.3.1) — [security advisory](https://github.com/advisories/GHSA-M6RX-H84Q-8R95)
- LangChain: RediSearch Filter Injection via Unescaped Tag/Text Values (1.1.1) — [security advisory](https://github.com/advisories/GHSA-5X6V-P487-7QH2)

### New APIs and options

- Adds native OpenAI-compatible events to streamEvents. (1.3.0) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.3.0)
- ChatGroq adds support for reasoning_effort. (1.1.5) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.1.5)
- Structured output now supports standard schemas. (1.1.4) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.1.4)
- Runnable traces now include package version metadata, making version information available in LangSmith trace metadata. (1.1.3) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.1.3)
- Adds an error class that contains partial output when a model invocation is aborted mid-stream. (1.0.4) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.0.4)

### Behavior and default changes

- Model profiles were updated, which may affect the profile information available to developers. (1.1.1) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.1.1)

### Fixes and other changes

- When aborted, stream() throws a regular AbortError because chunks have already been yielded to the caller. (1.0.4) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.0.4)
- Reasoning tokens are now properly elevated. (1.0.3) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.0.3)
- Fixed compatibility with the Node module resolution setting. (1.0.2) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/%40langchain/groq%401.0.2)

## @langchain/groq 0.3

### Security

- A fix for CVE 2025 12735 was included. (0.3.58) — [release notes](https://github.com/langchain-ai/langchainjs/releases/tag/0.3.58)