---
title: "Kimi k1.5: Scaling Reinforcement Learning with LLMs"
license: CC-BY-4.0
license_url: https://creativecommons.org/licenses/by/4.0/
license_note: 'paper is CC BY-NC-ND 4.0 (verified on arXiv abs page 2026-11-12); 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/2501-12599-kimi-k1.5.md
source_license: "CC BY-NC-ND 4.0"
attribution: 'Pointer by iinventedlight (AI-assisted). Source: Kimi Team and 95 coauthors (listed on the arXiv abs page), Kimi k1.5: Scaling Reinforcement Learning with LLMs (CC BY-NC-ND 4.0). This note links to the paper rather than reproducing it.'
author: 'iinventedlight (AI-assisted)'
date: 2026-11-12
status: published
approved_by: "Cole Withington, 2026-11-12 14:44 PT, Librarian chat widget t51s316: \"Yes, publish all 100 second-list arXiv notes\""
checked_on: 2026-11-12
as_of: 2026-11-12
confidence: medium
untested: 'Pointer note only. Paper not reproduced. Abs page and license checked 2026-11-12.'
category: arxiv
topic_area: 'ai-research: reasoning'
tags: arxiv, reasoning, reinforcement-learning, long-context, kimi
arxiv_id: 2501.12599
alias: 2501-12599-kimi-k1.5
type: pointer
format: pointer
sources:
  - id: S1
    title: "Kimi k1.5: Scaling Reinforcement Learning with LLMs"
    publisher: arXiv
    authors: "Kimi Team and 95 coauthors (listed on the arXiv abs page)"
    url: https://arxiv.org/abs/2501.12599
    version_read: "v4 (3 Jun 2025)"
    accessed: 2026-11-12
    license: "CC BY-NC-ND 4.0"
---
# Kimi k1.5: Scaling Reinforcement Learning with LLMs

## Summary
[Kimi k1.5: Scaling Reinforcement Learning with LLMs](https://arxiv.org/abs/2501.12599) (2025) is Moonshot's open report on scaling reinforcement learning for a long-context, multimodal reasoning model, published in the same period as DeepSeek-R1 and discussions of OpenAI's o1.

What it introduced: a relatively simple RL recipe that the authors say does not rely on Monte Carlo tree search, value functions, or process reward models. Long-context scaling and improved policy optimization are the main ingredients. They also describe long2short methods that use long chain-of-thought training to improve short chain-of-thought models. The report includes results on math, coding, and visual math benchmarks, including a comparison with o1 on several of those suites.

What it led to: one of the main public references for large-scale RL on reasoning models outside the DeepSeek-R1 line. Live notes: [2501-12948-deepseek-r1](https://iinventedlight.com/library/notes/2501-12948-deepseek-r1/) and [2402-03300-deepseekmath](https://iinventedlight.com/library/notes/2402-03300-deepseekmath/). Qwen2.5-Math is another pointer note in this batch.

Where to read it: https://arxiv.org/abs/2501.12599 (v4, last revised 3 Jun 2025).

The paper is posted under CC BY-NC-ND 4.0. 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: CC BY-NC-ND 4.0; 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).
