OpenAI o1 Contributions

OpenAI has released a detailed attribution list of the individuals and teams responsible for the developing the OpenAI o1 model. This publication serves as a formal record of the foundational, core, and supporting contributors who engineered the model and the specialized safety teams that ensured its robustness.

Personnel and Development Roles

OpenAI o1 was developed through a large-scale collaborative effort involving several tiers of contributors. The development process was led by a group of executive and supporting leadership, including Sam Altman, Mira Murati, and Greg Brockman.

Foundational and Core Contributors

The foundational contributors provided the primary technical architecture and research breakthroughs. This group includes individuals such as Ilya Sutskever, Jason Wei, and Noam Brown. Overall leadership for the project was provided by Jakub Pachocki, Jerry Tworek, Liam Fedus, Lukasz Kaiser, Mark Chen, Szymon Sidor, and Wojciech Zaremba.

Core contributors provided the essential engineering and research support necessary to bring the model to fruition, including researchers and engineers such as Alex Karpenko, Ananya Kumar, and David Dohan.

General Contributors

A broad team of contributors provided additional support across the rest of the development cycle, including personnel such as Aaron Jaech, Adam Lerer, and Reimar Leike.

Safety and Preparedness Framework

OpenAI implemented a comprehensive safety architecture for o1, dividing responsibilities between safety research, infrastructure, and evaluation teams to mitigate risks associated with the model's capabilities.

Safety Leadership and Research

Safety leadership was provided by Alex Beutel, Aleksander Madry, Irina Kofman, Joaquin Quiñonero Candela, Johannes Heidecke, Lilian Weng, and Mia Glaese. The o1 safety leads specifically included Boaz Barak, Eric Wallace, Jieqi Yu, Kai Xiao, Saachi Jain, Sam Toizer, Sandhini Agarwal, and Tejal Patwardhan.

Safety research was conducted by a team including Alec Helyar, Andrea Vallone, and Rahul Arora to refine the model's behavioral guardrails.

Safety Infrastructure and Evaluation

Safety infrastructure was managed by a team including Adam Richardson, Andrew Duberstein, and Vinnie Monaco, ensuring the technical systems for safety were operational.

To ensure the model met safety standards before release, OpenAI utilized both internal and external red teaming. Internal red teaming was conducted by Bowen Baker, Joost Huizinga, Karina Nguyen, and Wojciech Zaremba, while external red teaming was performed by Lama Ahmad, Michael Lampe, and Troy Peterson.

Preparedness Evaluations

Preparedness evaluations were handled by a team including Andy Applebaum, Elizabeth Proehl, and Jonathan Uesato, focusing on the model's potential for misuse or catastrophic risk.