Recursive Language Models Let LMs Call Themselves on Long Contexts

The RLM paper introduces a reasoning strategy where language models recursively call themselves on selected parts of the input context within a REPL. On the OOLONG benchmark, GPT-5-mini answered more than twice as many questions correctly as GPT-5.

2026-09-20 ~ 2026-09-20 · 2 related posts

• RLM explained: LMs recursively self-invoke in a REPL to tame long-context rot — gordic_aleksa · 2026-09-20
• RLM paper: GPT-5-mini-based RLM doubles GPT-5 accuracy on OOLONG, handles 10M+ tokens — gordic_aleksa · 2026-09-20