Open O1 is an open-source project designed to match the powerful capabilities of proprietary O1 models through open innovation. The project curates a set of O1-style thinking data to train the LLaMA and Qwen models, enhancing these smaller models' long-term reasoning and problem-solving abilities. As the Open O1 project progresses, we aim to further explore the possibilities of large language models, striving to create a model that not only achieves performance comparable to O1 but also excels in scalability during testing, making advanced AI capabilities accessible to all. Through community-driven development and a commitment to ethical practices, Open O1 will serve as a cornerstone for AI advancement, ensuring that the future development of technology is open and beneficial to everyone.