TY  - RPRT
TI  - Code-enabled language models can outperform reasoning models on diverse tasks
AU  - Cedegao E. Zhang
AU  - Cédric Colas
AU  - Gabriel Poesia
AU  - Joshua B. Tenenbaum
AU  - Jacob Andreas
PY  - 2025
UR  - https://arxiv.org/abs/2510.20909
ID  - 2510.20909
ER  -