RLM Library Adds Plug-and-Play Inference Support for Recursive Language Models
June 17, 2026
rlm is an open-source inference library for Recursive Language Models (RLMs) that supports multiple sandboxes and is designed as a drop-in addition to existing pipelines. No benchmark numbers or supported model list are provided.
HOW THIS AFFECTS YOU
●
builderYou can experiment with recursive LM inference patterns without building custom execution infrastructure.
●
researcherProvides a standardized sandbox environment for evaluating RLM architectures across different configurations.