gpt-oss-120b&gpt-oss-20b Model Card

/ Authors

OpenAI Sandhini Agarwal, L. Ahmad, Jason Ai, S. Altman, Andy Applebaum, Edwin Arbus, Rahul K. Arora, Yu Bai, Bowen Baker, Hai-Biao Bao

and 114 more authors

Boaz Barak, Ally Bennett, Tyler Bertao, N. Brett, E. Brevdo, Greg Brockman, Sébastien Bubeck, Cheng Chang, Kai Chen, Mark Chen, Enoch Cheung, Aidan Clark, Dan Cook, Marat Dukhan, C. Dvorak, K. Fives, Vlad Fomenko, T. Garipov, Kristian Georgiev, Mia Glaese, Tarun Gogineni, A. Goucher, L. Gross, K. Guzman, John Hallman, Jackie Hehir, Johannes Heidecke, Alec Helyar, Haitang Hu, Romain Huet, Jacob Huh, Saachi Jain, Z. Johnson, Chris Koch, Irina Kofman, Dominika Kundel, Jason Kwon, V. Kyrylov, Elaine Ya Le, Guillaume Leclerc, James Lennon, S. Lessans, Mario Lezcano-Casado, Yuanzhi Li, Zhuohan Li, Ji Lin, Jordan Liss, Lily Liu, Jiancheng Liu, K. Lu, Chris Lu, Z. Martinovic, Lindsay McCallum, Josh McGrath, Scott McKinney, A. McLaughlin, Song Mei, Steve Mostovoy, Tong Mu, Gideon Myles, Alexander Neitz, Alex Nichol, J. Pachocki, A. Paino, Dana Palmie, Ashley Pantuliano, Giambattista Parascandolo, Jongsoo Park, Leher Pathak, C. Paz, L. Peran, Dmitry Pimenov, Michelle Pokrass, Elizabeth Proehl, Huida Qiu, Gaby Raila, Filippo Raso, Hongyu Ren, K. Richardson, David Robinson, Bob Rotsted, Hadi Salman, Suvansh Sanjeev, Max Schwarzer, D. Sculley, Harshit S. Sikchi, Kendal Simon, Karan Singhal, Yang Song, Dane Stuckey, Zhiqing Sun, P. Tillet, Sam Toizer, Foivos Tsimpourlas, Nikhil Vyas, Eric Wallace, Xin Wang, Miles Wang, Olivia Watkins, Kevin Weil, Amy E. Wendling, Kevin Whinnery, Cedric Whitney, Hannah Wong, Lin Yang, Yu Yang, Michihiro Yasunaga, Kristen Ying, Wojciech Zaremba, Wenting Zhan, Cyril Zhang, B. Zhang, Eddie Zhang, Shengjia Zhao

/ Abstract

We present gpt-oss-120b and gpt-oss-20b, two open-weight reasoning models that push the frontier of accuracy and inference cost. The models use an efficient mixture-of-expert transformer architecture and are trained using large-scale distillation and reinforcement learning. We optimize the models to have strong agentic capabilities (deep research browsing, python tool use, and support for developer-provided functions), all while using a rendered chat format that enables clear instruction following and role delineation. Both models achieve strong results on benchmarks ranging from mathematics, coding, and safety. We release the model weights, inference implementations, tool environments, and tokenizers under an Apache 2.0 license to enable broad use and further research.