# Perplexity and NVIDIA unveil JOLT learning method

> The new post‑training technique lets a model teach itself, cutting token use by 6.5× on the MATH benchmark.

Oossa · 2026-10-22 · https://oossa.com/en/perplexity-and-nvidia-unveil-jolt-learning-method

Perplexity and NVIDIA researchers announced JOLT, a joint teacher‑student reinforcement learning method. The approach trains a single model to act as both teacher and student after the main training phase. On the MATH benchmark, JOLT reached the same accuracy as the GRPO method while using 6.5 times fewer training tokens. The result shows a way to improve efficiency without changing the model architecture.

## The facts

- JOLT matched GRPO’s accuracy on MATH with 6.5× fewer processed tokens.
- The announcement appeared in a MarkTechPost article on Oct 22, 2026.

## Why it matters

Researchers can fine‑tune models more cheaply, making high‑performance math reasoning more accessible to labs with limited compute budgets.

## Sources & references

1. [Perplexity and NVIDIA Researchers Introduce JOLT: a Joint Teacher-Student Reinforcement Learning Method](https://www.marktechpost.com/2026/10/21/perplexity-jolt-joint-on-policy-learning-teaching/) – MarkTechPost, 2026-10-22

Last updated: 2026-10-22