Code World Model Preparedness Report

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Song, Daniel, Ney, Peter, Menghini, Cristina, Ahmad, Faizan, Boyd, Aidan, Li, Nathaniel, Han, Ziwen, Testud, Jean-Christophe, Okabayashi, Saisuke, Ryan, Maeve, Miao, Jinpeng, Kwisaba, Hamza, Binder, Felix, Whitman, Spencer, Gust, Jim, Arcaute, Esteban, Kapil, Dhaval, Kahn, Jacob, Minhas, Ayaz, Goodman, Tristan, Deason, Lauren, Vaughan, Alexander, Zhao, Shengjia, Yue, Summer
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866909026535604224
author Song, Daniel
Ney, Peter
Menghini, Cristina
Ahmad, Faizan
Boyd, Aidan
Li, Nathaniel
Han, Ziwen
Testud, Jean-Christophe
Okabayashi, Saisuke
Ryan, Maeve
Miao, Jinpeng
Kwisaba, Hamza
Binder, Felix
Whitman, Spencer
Gust, Jim
Arcaute, Esteban
Kapil, Dhaval
Kahn, Jacob
Minhas, Ayaz
Goodman, Tristan
Deason, Lauren
Vaughan, Alexander
Zhao, Shengjia
Yue, Summer
author_facet Song, Daniel
Ney, Peter
Menghini, Cristina
Ahmad, Faizan
Boyd, Aidan
Li, Nathaniel
Han, Ziwen
Testud, Jean-Christophe
Okabayashi, Saisuke
Ryan, Maeve
Miao, Jinpeng
Kwisaba, Hamza
Binder, Felix
Whitman, Spencer
Gust, Jim
Arcaute, Esteban
Kapil, Dhaval
Kahn, Jacob
Minhas, Ayaz
Goodman, Tristan
Deason, Lauren
Vaughan, Alexander
Zhao, Shengjia
Yue, Summer
contents This report documents the preparedness assessment of Code World Model (CWM), a model for code generation and reasoning about code from Meta. We conducted pre-release testing across domains identified in our Frontier AI Framework as potentially presenting catastrophic risks, and also evaluated the model's misaligned propensities. Our assessment found that CWM does not pose additional frontier risks beyond those present in the current AI ecosystem. We therefore release it as an open-weight model.
format Preprint
id arxiv_https___arxiv_org_abs_2605_00932
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Code World Model Preparedness Report
Song, Daniel
Ney, Peter
Menghini, Cristina
Ahmad, Faizan
Boyd, Aidan
Li, Nathaniel
Han, Ziwen
Testud, Jean-Christophe
Okabayashi, Saisuke
Ryan, Maeve
Miao, Jinpeng
Kwisaba, Hamza
Binder, Felix
Whitman, Spencer
Gust, Jim
Arcaute, Esteban
Kapil, Dhaval
Kahn, Jacob
Minhas, Ayaz
Goodman, Tristan
Deason, Lauren
Vaughan, Alexander
Zhao, Shengjia
Yue, Summer
Software Engineering
Artificial Intelligence
This report documents the preparedness assessment of Code World Model (CWM), a model for code generation and reasoning about code from Meta. We conducted pre-release testing across domains identified in our Frontier AI Framework as potentially presenting catastrophic risks, and also evaluated the model's misaligned propensities. Our assessment found that CWM does not pose additional frontier risks beyond those present in the current AI ecosystem. We therefore release it as an open-weight model.
title Code World Model Preparedness Report
topic Software Engineering
Artificial Intelligence
url https://arxiv.org/abs/2605.00932