Exploring the System 1 Thinking Capability of Large Reasoning Models

Zhang, Wenyuan; Nie, Shuaiyi; Zhang, Xinghua; Zhang, Zefeng; Liu, Tingwen

doi:10.24963/ijcai.2026/676

Exploring the System 1 Thinking Capability of Large Reasoning Models

Wenyuan Zhang, Shuaiyi Nie, Xinghua Zhang, Zefeng Zhang, Tingwen Liu

Download BibTeX Cite

Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence. Main Track. Pages 6073-6081.

https://doi.org/10.24963/ijcai.2026/676

Keywords

Natural Language Processing: Resources and evaluation
Natural Language Processing: Other

Abstract

This paper explores the system 1 thinking capability of Large Reasoning Models (LRMs), the intuitive ability to respond efficiently with minimal token usage. While existing LRMs rely on long-chain reasoning and excel at complex tasks, their system 1 thinking ability remains largely underexplored. This capability is essential as it reflects models' difficulty awareness and reasoning efficiency, both critical for real-world applications. We propose S1-Bench, a multi-domain, multilingual benchmark comprising model-simple system 1 questions. Our investigation of 28 LRMs reveals under-accuracy and inefficiency on system 1 problems. We find existing efficient reasoning methods either generalize poorly to simple questions or sacrifice performance for efficiency. Further exploration uncovers LRMs' early difficulty awareness accompanied by lower confidence, and shows that problem difficulty is implicitly encoded in hidden states. Code and the extended version of the paper with appendices are available at https://github.com/WYRipple/S1_Bench.

Wenyuan Zhang, Shuaiyi Nie, Xinghua Zhang, Zefeng Zhang, and Tingwen Liu. 2026. Exploring the System 1 Thinking Capability of Large Reasoning Models. In Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26. IJCAI, 6073–6081. https://doi.org/10.24963/ijcai.2026/676

@inproceedings{ijcai2026p676,
  title     = {Exploring the System 1 Thinking Capability of Large Reasoning Models},
  author    = {Zhang, Wenyuan and Nie, Shuaiyi and Zhang, Xinghua and Zhang, Zefeng and Liu, Tingwen},
  booktitle = {Proceedings of the Thirty-Fifth International Joint Conference on
               Artificial Intelligence, {IJCAI-26}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Diego Calvanese},
  pages     = {6073--6081},
  year      = {2026},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2026/676},
  url       = {https://doi.org/10.24963/ijcai.2026/676},
}

TY  - CPAPER
TI  - Exploring the System 1 Thinking Capability of Large Reasoning Models
AU  - Zhang, Wenyuan
AU  - Nie, Shuaiyi
AU  - Zhang, Xinghua
AU  - Zhang, Zefeng
AU  - Liu, Tingwen
T2  - Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2026
SP  - 6073
EP  - 6081
DO  - 10.24963/ijcai.2026/676
UR  - https://doi.org/10.24963/ijcai.2026/676
ER  -