---
type: "post"
id: "perplexity-and-nvidia-released-jolt-a-joint-teacher-0cb0aa"
slug: "perplexity-and-nvidia-released-jolt-a-joint-teacher-0cb0aa"
title: "Perplexity and NVIDIA released JOLT , a joint teacher-student self-distillation post-training method."
description: "Perplexity and NVIDIA have described JOLT, short for Joint On-policy Learning and Teaching, a post-training method for teacher-student self-distillation."
retrieval_nugget: "Perplexity and NVIDIA have described JOLT, short for Joint On-policy Learning and Teaching, a post-training method for teacher-student self-distillation."
published_at: "2026-10-22"
updated_at: "2026-10-22"
record_date: "2026-10-22"
date_kind: "published_at"
topics: ["models"]
entities: ["Perplexity","NVIDIA","JOLT","Qwen3","DeepSeek-V2-Lite"]
related_patterns: []
novelty_label: "new_fact"
what_changed: "Perplexity and NVIDIA released JOLT, a joint teacher–student self-distillation post-training method, distinct from NVIDIA Research’s earlier offline-rollout comparison."
prior_refs: ["memory:76428e7c-e1ec-8d85-acf1-182dc1d5a751","memory:e4cab375-c6e9-8cff-adb0-084839023f6a","memory:73aa2d11-5af0-87ed-a2c1-287ef3d1bb6d"]
source_url: "https://www.perplexity.ai/hub/blog/a-good-teacher-meets-the-student-where-they-are"
source_urls: ["https://www.perplexity.ai/hub/blog/a-good-teacher-meets-the-student-where-they-are"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:perplexity-and-nvidia-released-jolt-a-joint-teacher-0cb0aa"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/perplexity-and-nvidia-released-jolt-a-joint-teacher-0cb0aa/","markdown":"https://newruntime.com/posts/perplexity-and-nvidia-released-jolt-a-joint-teacher-0cb0aa.md","json":"https://newruntime.com/posts/perplexity-and-nvidia-released-jolt-a-joint-teacher-0cb0aa.json"}
---

# Perplexity and NVIDIA released JOLT , a joint teacher-student self-distillation post-training method.

## Retrieval answer

Perplexity and NVIDIA have described JOLT, short for Joint On-policy Learning and Teaching, a post-training method for teacher-student self-distillation.

Perplexity and NVIDIA have described JOLT, short for Joint On-policy Learning and Teaching, a post-training method for teacher-student self-distillation. The work targets reinforcement-learning settings in which outcome rewards are sparse, particularly for long reasoning and agentic trajectories.

JOLT uses one model in two roles. The teacher receives privileged information, such as verified solutions, while the student does not. Both roles are trained jointly: the student follows a distillation objective, and the teacher is optimized to provide feedback that moves the student in a direction aligned with the student's own RL reward.

The authors report improvements in model performance and training efficiency on mathematical reasoning, coding, and agentic tasks. The accompanying material identifies Qwen3 and DeepSeek-V2-Lite as base-model families used in evaluation. It does not establish that the results generalize to other model families, datasets, workloads, or production deployments, nor does the supplied source provide an independent replication.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Models](https://newruntime.com/topics/models.md): Explore the models topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)