LLMs Research

LLMs Research

Recursive Language Models: Treating the Prompt as Code

Recursive Language Models store the prompt as a variable and let the model write code to read it

LLMs Research's avatar
LLMs Research
Feb 23, 2026
∙ Paid

Have you felt this, that your LLM suddenly becomes dumb and starts hallucinating? We all thought it was a context problem. The entire AI industry has been trying to fix it by making the context window bigger. GPT-5 has a 272K-token window. Gemini advertises 2M tokens. But it doesn’t work.

When researchers from MIT tested GPT-5 on BrowseComp+, a benchmark…

User's avatar

Continue reading this post for free, courtesy of LLMs Research.

Or purchase a paid subscription.
© 2026 LLMs Research Inc. · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for great culture