In a move toward fostering collaboration and innovation in the AI community, Elon Musk’s xAI has open-sourced the base code of its Grok AI model, without including any training code. Termed as the “314 billion parameter Mixture-of-Expert model,” the code is now available on GitHub, under the Apache License 2.0, allowing for commercial use cases.

In a recent blog post, xAI clarified that the Grok model, labeled as Grok-1, was not specifically tuned for any particular application, such as conversational use. Although originally trained on a “custom” stack, the company did not disclose further details regarding its training data. Notably, the open-source release of Grok does not include connections to the X social network, contrasting with its earlier chatbot version accessible to Premium+ users.

The decision to open-source Grok follows Musk's announcement last week on X, signaling xAI’s commitment to fostering transparency and collaboration in the AI landscape. Despite ongoing legal battles with OpenAI, including a recent lawsuit over alleged breaches of nonprofit AI goals, Musk remains dedicated to advancing AI research and development. As the industry witnesses a surge in open-source AI models, Grok’s availability presents new opportunities for AI tool makers to integrate and innovate in their solutions, promising further advancements in conversational AI and beyond.