People often use these terms interchangeably, but they don’t mean quite the same thing. An open-weight model can also be open-source, but publishing the weights alone doesn’t qualify under the Open Source Initiative’s definition.
The Open Source Initiative (OSI) has already applied this in the context of AI, with their open source AI definition (OSAID) according to which an open source AI system is one that enables four key freedoms.
The OSI recently launched an Open Source AI Fellowship, appointing Gabriel Toscano as its first fellow to help build consensus around what constitutes open source AI. The two-year program will explore potential revisions to the definition, while OSI also plans a series of community discussions, or “open source salons,” over the next two years.
The OSI has its own definition of open source AI: the Open Source AI Definition (OSAID 1.0). It requires model parameters, including weights, to be made available under OSI-approved terms, but does not prescribe a specific legal mechanism for doing so.
The Open Source Initiative’s Open Source AI Definition requires opening the parameters, the complete training and inference code, and enough information about the data to rebuild a substantially equivalent model.
Under the Open Source Initiative’s Open Source AI Definition, truly open-source AI requires more than downloadable weights. Users need broad rights to use, study, modify and share the system, along with access to parameters, relevant code and sufficient information about the training data and process to meaningfully modify the system. Many popular models commonly described as open source therefore fit more accurately into the broader open-weight category.
OSI released its first official draft of the OSAID in 2025. The OSAID closely tracks the OSI’s criteria for open source software.
As noted by The Open Source Initiative (OSI), a nonprofit, open source models are fully transparent. Not only can users download the models to run locally on their machines, but their full training code and training datasets (or at least data sources, if the datasets aren’t legally permitted to be copied) are also available for examination.
According to the Open Source Initiative (OSI), a truly open-source model also releases the data it was trained on, along with other information allowing those models to be studied, inspected, used, modified and freely distributed.
“We welcome Hugging Face’s work on the benefits of ‘open’ models, and we need more of it from everyone who has ever built on open source,” said Katie Steen-James, a senior U.S. policy manager at the nonprofit Open Source Initiative, which promotes open-source software.
The difference between open-source AI and weights, according to the Open Source Initiative (OSI), is that open weights usually means the trained weights are available, but not necessarily the full training data or source code. The two terms are frequently conflated. The OSI itself declared, “Every AI system in the headlines today, whether proprietary or open source, exists because researchers shared their work openly.” Without open source, there is no AI.
The Open Source Initiative’s definition of a fully open-source AI model includes the training data as a key element. Some developers wonder, however, how feasible it is to distribute the enormous datasets required.
