Can Large Language Models Reason and Optimize Under Constraints?

Fabien Bernier, Salah Ghamizi, Pantelis Dogoulis, Maxime Cordy · 2026

Read Can Large Language Models Reason and Optimize Under Constraints? by Fabien Bernier, Salah Ghamizi, Pantelis Dogoulis, Maxime Cordy (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/arxiv-2603-23004.txt.

Description

Bernier and co-authors test whether large language models can reason and optimize under constraints. Their testbed is the optimal power flow problem from power grids. Solving it needs reasoning, structured input handling, arithmetic, and constrained optimization at once. Top models fail most of the tasks, they find. Reasoning-focused models still fail in the hardest settings. The authors present the setup as a strict test for future assistants. Real grid optimization is the long-term target.

Catalog record

CREATORFabien Bernier, Salah Ghamizi, Pantelis Dogoulis, Maxime Cordy
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34391
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.23004

Browse the full library · About the Tower of Babel