Performance of Reasoning and Comparator Large Language Models on Nephrology Multiple-choice Questions
Fumiya Kitano, Mamoru Masaki, Daisuke Ichikawa, Yugo Shibagaki, Ryunosuke Noda · 2026
Read Performance of Reasoning and Comparator Large Language Models on Nephrology Multiple-choice Questions by Fumiya Kitano, Mamoru Masaki, Daisuke Ichikawa, Yugo Shibagaki, Ryunosuke Noda (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13639661.txt.
Description
Fumiya Kitano and four co-authors test reasoning language models against plain ones on kidney medicine exam questions, in a study from 2026. Reasoning models shine in general medicine, but whether their edge depends on question type in a specialty was unclear. Board renewal questions give a fair test.
Two hundred nine Japanese nephrology questions, some with images, went to four models from two makers. The reasoning models scored about eighty-eight and eighty-four percent, clearly beating comparators near seventy and sixty-three percent. Odds strongly favored reasoning across question types, from recall to problem solving. Statistical models confirmed the gap was not chance.
Catalog record
| CREATOR | Fumiya Kitano, Mamoru Masaki, Daisuke Ichikawa, Yugo Shibagaki, Ryunosuke Noda |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |