---
title: 'OpenAI o1 System Card'
license: CC-BY-4.0
license_url: https://creativecommons.org/licenses/by/4.0/
license_note: 'paper is CC BY 4.0 (verified on arXiv abs page 2026-11-12; icon by-4.0.png, http://creativecommons.org/licenses/by/4.0/); note is our own summary'
source: https://iinventedlight.com/library/notes/2412-16720-o1-system-card.md
source_license: 'CC BY 4.0'
attribution: 'Summary by iinventedlight (AI-assisted). Source: OpenAI and 264 coauthors (listed on the arXiv abs page), OpenAI o1 System Card (CC BY 4.0).'
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: 'Not reproduced. Abs page and HTML key sections read 2026-11-12. Named code URLs checked for HTTP 200; code not run.'
category: arxiv
topic_area: 'ai-research: reasoning'
tags: arxiv, reasoning, alignment, system-card
arxiv_id: 2412.16720
alias: 2412.16720-o1-system-card
sources:
  - id: S1
    title: 'OpenAI o1 System Card'
    publisher: arXiv
    authors: 'OpenAI and 264 coauthors (listed on the arXiv abs page)'
    url: https://arxiv.org/abs/2412.16720
    version_read: 'v2 (30 Apr 2026)'
    accessed: 2026-11-12
    license: 'CC BY 4.0'
---
# OpenAI o1 System Card

## Summary
The o1 model series is trained with large-scale reinforcement learning to reason using chain of thought. The authors say these reasoning capabilities provide new avenues for improving safety and robustness, including by reasoning about safety policies in context through deliberative alignment. This report outlines the safety work carried out for the OpenAI o1 and OpenAI o1-mini models, including safety evaluations, external red teaming, and Preparedness Framework evaluations. [S1]

## Claims
- id: C1
  claim: 'The o1 series is trained with large-scale reinforcement learning to reason using chain of thought.'
  source: S1
  evidence: "trained with large-scale reinforcement learning to reason using chain of thought"
- id: C2
  claim: 'The report covers safety work for OpenAI o1 and OpenAI o1-mini.'
  source: S1
  evidence: "safety work carried out for the OpenAI o1 and OpenAI o1-mini models"
- id: C3
  claim: 'That work includes safety evaluations, external red teaming, and Preparedness Framework evaluations.'
  source: S1
  evidence: "including safety evaluations, external red teaming, and Preparedness Framework evaluations"

## Verification
- Code available: No official model or training-code URL was named as a release of this paper on the abs page or HTML. The HTML points to https://www.github.com/openai/simple-evals, which returned HTTP 200 and redirected to https://github.com/openai/simple-evals on 2026-11-12. That is an evaluation repo, not the o1 model. Not run.
- Reproduced: not reproduced (2026-11-12). Code not run.
- Related work: Live notes [2408-03314-test-time-compute](https://iinventedlight.com/library/notes/2408-03314-test-time-compute/) and [2501-12948-deepseek-r1](https://iinventedlight.com/library/notes/2501-12948-deepseek-r1/). This is the paper-verification note for the OpenAI o1 system card itself.
- Limits: This is a company system card, not an independent audit. Headline safety claims are author-reported. This note follows the v2 text last revised 30 Apr 2026 and does not restate every evaluation table.
- License: CC BY 4.0 on the abs page (2026-11-12; icon by-4.0.png; http://creativecommons.org/licenses/by/4.0/). Submitted 21 Dec 2024 (v1); last revised 30 Apr 2026 (v2). Sources: [S1] https://arxiv.org/abs/2412.16720, v2 (30 Apr 2026). checked_on: 2026-11-12.

## When to apply
Use when asking how OpenAI described o1-style chain-of-thought training and the safety evaluations around it, not as a substitute for an independent eval.

## Facts
- id: F1
  claim: 'The o1 model series is trained with large-scale reinforcement learning to reason using chain of thought.'
  confidence: medium
  date: 2026-11-12
  sources: [S1]
  license: CC BY 4.0
- id: F2
  claim: 'The report outlines safety work for the OpenAI o1 and OpenAI o1-mini models.'
  confidence: medium
  date: 2026-11-12
  sources: [S1]
  license: CC BY 4.0
- id: F3
  claim: 'The described work includes safety evaluations, external red teaming, and Preparedness Framework evaluations.'
  confidence: medium
  date: 2026-11-12
  sources: [S1]
  license: CC BY 4.0

## Sources
- [S1] arXiv 2412.16720, https://arxiv.org/abs/2412.16720. License: CC BY 4.0. v2 (30 Apr 2026). Accessed 2026-11-12.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Summary by iinventedlight (AI-assisted). Paper is CC BY 4.0.

---

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