---
title: 'DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning'
license: CC-BY-4.0
license_url: https://creativecommons.org/licenses/by/4.0/
license_note: 'paper is arXiv non-exclusive distribution license (verified on arXiv abs page 2026-11-11); this pointer note is our own words and is CC BY 4.0; we link to the paper rather than reproduce it'
source: https://iinventedlight.com/library/notes/2501-12948-deepseek-r1.md
source_license: 'arXiv non-exclusive distribution license'
attribution: 'Pointer by iinventedlight (AI-assisted). Source: DeepSeek-AI and coauthors listed on the arXiv abs page, DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning (arXiv non-exclusive distribution license). This note links to the paper rather than reproducing it.'
author: 'iinventedlight (AI-assisted)'
date: 2026-11-11
status: published
approved_by: "Cole Withington, 2026-11-11 20:58 PT, Grok Bot chat widget t1717s7407: \"Publish all 100 including the LLMs Encode Clinical Knowledge note\""
checked_on: 2026-11-11
as_of: 2026-11-11
confidence: medium
untested: 'Pointer note only. Paper not reproduced. Abs page and license checked 2026-11-11.'
category: arxiv
topic_area: 'ai-research: reasoning'
tags: arxiv, reasoning, reinforcement-learning, open-weights
arxiv_id: 2501.12948
alias: 2501.12948-deepseek-r1
type: pointer
format: pointer
sources:
  - id: S1
    title: 'DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning'
    publisher: arXiv
    authors: 'DeepSeek-AI and coauthors listed on the arXiv abs page'
    url: https://arxiv.org/abs/2501.12948
    version_read: 'v2 (4 Jan 2026)'
    accessed: 2026-11-11
    license: 'arXiv non-exclusive distribution license'
---
# DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

## Summary
[DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://arxiv.org/abs/2501.12948) (2025) is the public write-up of DeepSeek-R1. It showed that large-scale reinforcement learning, with a carefully chosen reward setup, can push a language model toward longer, more reliable reasoning traces, and it did so with open weights and a recipe other labs could study.

What it introduced: the DeepSeek-R1 training path, including a reinforcement-learning first stage (R1-Zero) and a later pipeline that combines that signal with supervised data, plus a released family of models.

What it led to: a wave of open reasoning models and a shift toward RL-at-scale as a post-training method, not only supervised instruction data. For the earlier preference-training background, see the InstructGPT pointer in this batch and the live notes [2204-05862-helpful-harmless-rlhf](https://iinventedlight.com/library/notes/2204-05862-helpful-harmless-rlhf/) and [2305-18290-direct-preference-optimization](https://iinventedlight.com/library/notes/2305-18290-direct-preference-optimization/).

Where to read it: https://arxiv.org/abs/2501.12948 (v2, last revised 4 Jan 2026).

The paper is posted on arXiv under the arXiv non-exclusive distribution license. This note links to it rather than reproducing it. The note itself is CC BY 4.0.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Pointer by iinventedlight (AI-assisted). Paper license: arXiv non-exclusive distribution; this note links to it and does not reproduce it.

---

Agents: [/agents.md](/agents.md) | [/.well-known/agent-library.json](/.well-known/agent-library.json).