AI ChronicleAI 纪年 Full timeline →

Model release · curated entry

DeepSeek-R1 brings open-weight reasoning near the closed frontier

DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

Date
January 20, 2025
Category
Model release
Organization / Model
DeepSeek
Evidence
2 sources · 2 evidence families

R1 used reinforcement learning to elicit long chain-of-thought reasoning and opened its weights under the MIT licence. It approached closed frontier models on several reasoning benchmarks while cutting serving cost sharply.

Curator's note: Cross-checked against the paper and the Hugging Face model page.

Curated sources · 2

  1. Paper preprints · arXiv2025-01-20

    DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

    arxiv.org

  2. Open model hub · Hugging Face2025-01-20

    DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

    huggingface.co

Cite this page

This page is a curated static snapshot and can be cited directly. Preferred format GB/T 7714—2025 (electronic resource [EB/OL]); the other 5 styles open below, together with the machine-readable files.

GB/T 7714—2025 (electronic resource [EB/OL])

AI 纪年. DeepSeek-R1 发布，开放权重推理模型接近闭源前沿[EB/OL]. (2025-01-20)[2026-11-03]. https://aihist.net/e/deepseek-r1-2025.html.

APA 7 · American Psychological Association

AI 纪年. (2025, January 20). DeepSeek-R1 发布，开放权重推理模型接近闭源前沿 [Web page]. https://aihist.net/e/deepseek-r1-2025.html

MLA 9 · Modern Language Association

"DeepSeek-R1 发布，开放权重推理模型接近闭源前沿." AI 纪年, 20 Jan. 2025, aihist.net/e/deepseek-r1-2025.html. Accessed 3 Nov. 2026.

Chicago 17 · notes-bibliography

AI 纪年. "DeepSeek-R1 发布，开放权重推理模型接近闭源前沿." Published Jan. 20, 2025. Accessed Nov. 3, 2026. https://aihist.net/e/deepseek-r1-2025.html.

IEEE · citation-reference list

"DeepSeek-R1 发布，开放权重推理模型接近闭源前沿," AI 纪年, 20 Jan. 2025. [Online]. Available: https://aihist.net/e/deepseek-r1-2025.html. [Accessed: 3 Nov. 2026].

ACM · author-date variant

AI 纪年. January 20, 2025. DeepSeek-R1 发布，开放权重推理模型接近闭源前沿. Retrieved November 3, 2026 from https://aihist.net/e/deepseek-r1-2025.html.

All six styles come from one template and can be compared word for word. [引用日期], Accessed, urldate and Y2 all carry this page's build date, 2026-11-03 — a static page is an immutable snapshot, so substitute the date you actually consulted it. The address in every style is the archive's canonical page: this English page translates the same snapshot, and the work keeps its name, so one record cites both. For one-click copying use the same entry in the timeline, which takes the day you opened it. Expand the other styles before printing.