(How) Do reasoning models reason?

Subbarao Kambhampati; Kaya Stechly; Karthik Valmeekam

doi:10.1111/nyas.15339

(How) Do reasoning models reason?

Subbarao Kambhampati, Kaya Stechly, Karthik Valmeekam

Research output: Contribution to journal › Comment/debate › peer-review

1 Scopus citations

Abstract

We provide a broad unifying perspective on the recent breed of large reasoning models such as OpenAI o1 and DeepSeek R1, including their promise, sources of power, misconceptions, and limitations.

Original language	English (US)
Pages (from-to)	33-40
Number of pages	8
Journal	Annals of the New York Academy of Sciences
Volume	1547
Issue number	1
DOIs	https://doi.org/10.1111/nyas.15339
State	Published - May 2025
Externally published	Yes

Keywords

large language models
planning
reasoning
reinforcement learning post-training
test-time inference

ASJC Scopus subject areas

General Neuroscience
General Biochemistry, Genetics and Molecular Biology
History and Philosophy of Science

Access to Document

10.1111/nyas.15339

Cite this

@article{618ca06c3607441885df4523e689ac22,

title = "(How) Do reasoning models reason?",

abstract = "We provide a broad unifying perspective on the recent breed of large reasoning models such as OpenAI o1 and DeepSeek R1, including their promise, sources of power, misconceptions, and limitations.",

keywords = "large language models, planning, reasoning, reinforcement learning post-training, test-time inference",

author = "Subbarao Kambhampati and Kaya Stechly and Karthik Valmeekam",

note = "Publisher Copyright: {\textcopyright} 2025 The New York Academy of Sciences.",

year = "2025",

month = may,

doi = "10.1111/nyas.15339",

language = "English (US)",

volume = "1547",

pages = "33--40",

journal = "Annals of the New York Academy of Sciences",

issn = "0077-8923",