---
format: "aidr-story-markdown/v1"
id: "339b8d911d2d7a0fea0e5be72c09b9f5e65482d5a488b9ee1d49febc5d87a4b8"
canonical_url: "https://aidr.today/339b8d91?lang=en"
title: "Perplexity and NVIDIA Researchers Introduce JOLT: a Joint Teacher-Student Reinforcement Learning Method"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-22T04:56:31.000Z"
category: "Research"
topics: ["nvidia","llm"]
source_urls: ["https://www.marktechpost.com/2026/10/21/perplexity-jolt-joint-on-policy-learning-teaching/"]
summary: "Perplexity and NVIDIA introduced JOLT, a post-training method that trains one model as both teacher and student. On MATH it matched GRPO’s accuracy with 6.5x fewer processed training tokens."
---

# Perplexity and NVIDIA Researchers Introduce JOLT: a Joint Teacher\-Student Reinforcement Learning Method

> [Open the canonical story](<https://aidr.today/339b8d91?lang=en>)

**Published:** 2026-10-22T04:56:31.000Z
**Category:** Research
**Topics:** nvidia, llm

## Summary

Perplexity and NVIDIA introduced JOLT, a post\-training method that trains one model as both teacher and student\. On MATH it matched GRPO’s accuracy with 6\.5x fewer processed training tokens\.

## Sources

- [Story source](<https://www.marktechpost.com/2026/10/21/perplexity-jolt-joint-on-policy-learning-teaching/>)