Open models by OpenAI
Advanced open-weight reasoning models to customize for any use case and run anywhere.
Permissive license
These models are supported by the Apache 2.0 license. Build freely without worrying about copyleft restrictions or patent risk—whether you're experimenting, customizing, or deploying commercially.
Designed for agentic tasks
Leverage powerful instruction following and tool use within the chain-of-thought, including web search and Python code execution.
Deeply customizable
Adjust the reasoning effort to low, medium, or high. Plus, customize the models to adapt to your use case with full-parameter fine-tuning.
Full chain-of-thought
Access the full chain-of-thought for easier debugging and higher trust in model outputs.
Interactive demo
We built a simple playground where developers can try both models in the browser.
Model performance
gpt-oss-120b | gpt-oss-20b | OpenAI o3 | OpenAI o4-mini | |
Reasoning & knowledge | ||||
MMLU | 90.0 | 85.3 | 93.4 | 93.0 |
GPQA Diamond | 80.1 | 71.5 | 83.3 | 81.4 |
Humanity’s Last Exam | 19.0 | 17.3 | 24.9 | 17.7 |
Competition math | ||||
AIME 2024 | 96.6 | 96.0 | 95.2 | 98.7 |
AIME 2025 | 97.9 | 98.7 | 98.4 | 99.5 |
Advancing safety standards for open models
Safety is foundational to our open models
Each model has completed thorough safety training and evaluation to help developers keep users safe.
Comprehensive safety testing
We rigorously tested a maliciously fine-tuned version of gpt-oss-120b under our Preparedness Framework, and found that it doesn’t reach high capability levels. These training and testing methods were reviewed and informed by external safety experts, and mark a meaningful advancement in open model safety standards.
Our partners
We’re working with leading deployment and hardware companies to offer these models to the open-source community.
Deployment partners
Hardware partners
Resources
We want to hear from you
Share feedback and feature requests that may guide future open models. Please note: you won’t receive a direct response. For support and discussion, visit our Hugging Face Community.