open-weight model
In plain wordsAn AI model whose trained internal numbers are published for anyone to download and run on their own computers, even if the data and code used to make it stay private.
A trained model whose parametersweights, n.: The learned numbers inside a neural network that set how strongly each part influences the next. The finished set of weights is, in effect, the trained model itself. are released for public download, so that anyone can run it on their own hardware, inspect its internals, fine-tunefine-tuning, n.: Taking an AI model that has already learned a lot and training it a little more on a smaller, focused set of examples so it gets better at one job or behaves a certain way. or quantizequantization, n.: Storing an AI model's numbers with fewer digits, like rounding prices to the nearest dollar, so the model takes less memory and runs faster, usually with only a small loss of accuracy. it, and build products on it, subject to its licensemodel license, n.: The legal terms under which an AI model may be downloaded and used — for example, whether businesses may use it, whether changes may be shared, and which uses are banned.. Releases range from weights under permissive licenses such as Apache 2.0 or MIT (Mistral 7B, DeepSeek-R1DeepSeek-R1, prop. n.: A free, downloadable AI reasoning model released by the Chinese company DeepSeek in January 2025. It matched leading American models at math and coding for a fraction of the reported cost, shaking the industry., OpenAI's gpt-oss) to weights under custom licenses with use restrictions (Meta's LlamaLlama, prop. n.: A family of AI language models from Meta, first released in 2023, whose trained settings can be downloaded and run by anyone, making them the foundation for thousands of other AI projects. models). 'Open-weight' is narrower than open-source AIopen-source AI, n.: AI released, ideally with its code and details of its training data and not just the finished model, so that anyone may use, inspect, alter, share and even rebuild it. How open is open enough is debated.: training data, data-processing code and training recipes are often withheld, so the model can be used and adapted but not fully reproduced or audited. Supporters cite research access, privacy through local deployment, competition, independent scrutiny and technological self-reliance; critics cite that a release cannot be recalled, that safety training can be removed by fine-tuning, and possible misuse for cyber or biological harm. Policy analyses therefore weigh the marginal risk a release adds beyond closed models and existing tools.
Because it is an open-weight model, the hospital runs it entirely on its own servers.
Origin
Researchers have long published trained weights, and OpenAI's staged release of GPT-2GPT-2, prop. n.: A 2019 OpenAI language model that wrote surprisingly fluent paragraphs. OpenAI at first withheld the full version, saying it could be misused for fake news, which made it famous. in 2019 made release itself a policy question. The distinct label 'open-weight' spread in 2023–2024, as debates over models such as Llama 2 separated publishing weights from open-source software; the US NTIA's July 2024 report on models with 'widely available model weights' framed the policy debate.
Usage note
Often called 'open-source models', a usage the Open Source Initiative and others dispute when licenses restrict use or training data is withheld. Under the EU AI ActEU AI Act, prop. n.: The European Union's law on artificial intelligence, the first broad AI law anywhere: it bans some uses, sets strict checks for risky ones and puts duties on the makers of the biggest AI models., general-purpose modelsgeneral-purpose AI model, n.: The EU's legal name for the large, adaptable AI models behind chatbots and similar tools: models that can turn their hand to many different jobs and be built into many other products. released under free and open-source licenses with public weights are exempt from some documentation duties unless they pose systemic risk.
Related
- open-source AIopen-source AI, n.: AI released, ideally with its code and details of its training data and not just the finished model, so that anyone may use, inspect, alter, share and even rebuild it. How open is open enough is debated.
- proprietary modelproprietary model, n.: An AI model kept private by the company that made it; outsiders can use it only through the company's apps or paid online services and can never download it.
- model licensemodel license, n.: The legal terms under which an AI model may be downloaded and used — for example, whether businesses may use it, whether changes may be shared, and which uses are banned.
- Hugging FaceHugging Face, prop. n.: A company whose website is the main public library for sharing AI models and datasets — often called the GitHub of machine learning — and which makes widely used free AI software.
- dual-use foundation modeldual-use foundation model, n.: A US legal term for a general-purpose AI model that serves everyday uses but is capable enough that it could help someone build weapons, mount cyberattacks or evade human control.
- LlamaLlama, prop. n.: A family of AI language models from Meta, first released in 2023, whose trained settings can be downloaded and run by anyone, making them the foundation for thousands of other AI projects.
- fine-tuningfine-tuning, n.: Taking an AI model that has already learned a lot and training it a little more on a smaller, focused set of examples so it gets better at one job or behaves a certain way.
- proprietary modelproprietary model, n.: An AI model kept private by the company that made it; outsiders can use it only through the company's apps or paid online services and can never download it.
Sources
- National Telecommunications and Information Administration (2024). Dual-Use Foundation Models with Widely Available Model Weights. US Department of Commerce, July 2024.
Mentioned in
America's AI Action PlanAmerica's AI Action Plan, prop. n.: The Trump administration's 2025 strategy for winning a global AI race: cutting regulation, building data centers and power plants quickly, and exporting American AI technology to allies., DeepSeekDeepSeek, prop. n.: A Chinese artificial intelligence company whose free, cheaply trained models startled the world in January 2025 by nearly matching the best American systems., frontier modelfrontier model, n.: One of the most capable AI models in existence at a given time, built by a handful of leading labs at great cost, and the main target of AI safety rules., GeminiGemini, prop. n.: Google's family of AI models and its chatbot assistant, launched in December 2023. Gemini models handle text, images, audio and video and power many Google products., Hugging FaceHugging Face, prop. n.: A company whose website is the main public library for sharing AI models and datasets — often called the GitHub of machine learning — and which makes widely used free AI software., LlamaLlama, prop. n.: A family of AI language models from Meta, first released in 2023, whose trained settings can be downloaded and run by anyone, making them the foundation for thousands of other AI projects., Meta AIMeta AI, prop. n.: The artificial intelligence research group of Meta, the company behind Facebook and Instagram, known for the PyTorch software and the Llama models. It is also the name of Meta's chatbot., misusemisuse, n.: People deliberately using AI to cause harm — for scams, harassment, propaganda, hacking or weapons — as opposed to harm from AI making mistakes or pursuing goals of its own., model APImodel API, n.: A web service through which programs send requests to an AI model run by a company and receive its answers, usually paying per use instead of running the model themselves., on-device AIon-device AI, n.: AI that runs directly on your own phone, laptop, car or headset rather than on a company's distant servers, so it can work offline, respond instantly and keep your data on the device., open-source AIopen-source AI, n.: AI released, ideally with its code and details of its training data and not just the finished model, so that anyone may use, inspect, alter, share and even rebuild it. How open is open enough is debated., proprietary modelproprietary model, n.: An AI model kept private by the company that made it; outsiders can use it only through the company's apps or paid online services and can never download it., refusalrefusal, n.: When an AI declines to do what it was asked — 'I can't help with that' — usually because the request seems harmful or breaks the rules it was trained to follow., Stable DiffusionStable Diffusion, prop. n.: A widely used image-generating AI whose model files were released for anyone to download, so people could run it on their own computers and adapt it., Yann LeCunYann LeCun, prop. n.: A French-American computer scientist who pioneered the neural networks that recognize images, led Meta's AI research for over a decade, and argues that chatbots alone will not reach human-level AI., activation steeringactivation steering, n.: Changing an AI model's behavior by adjusting the internal numbers it computes while it runs — like turning a hidden dial — instead of rewording the prompt or retraining the model., AI engineeringAI engineering, n.: The craft of building useful, reliable products on top of AI models — choosing models, writing prompts, connecting data and tools, testing quality and running it all safely at scale., AI modelAI model, n.: The trained core of an AI: a large file of learned numbers plus the recipe for using them, which turns an input such as a question or a photo into an output., AlibabaAlibaba, prop. n.: A Chinese e-commerce and cloud-computing giant whose Qwen family of AI models, mostly free to download, is among the most widely used in the world., Allen Institute for AIAllen Institute for AI, prop. n.: Paul Allen, a co-founder of Microsoft, created this nonprofit Seattle AI lab in 2014. It is known for releasing models together with everything used to build them, and for the Semantic Scholar paper search engine., black boxblack box, n.: A system whose inner workings are hidden or too complex to understand: we can see what goes in and what comes out, but not why it gave a particular answer., Center for AI Standards and InnovationCenter for AI Standards and Innovation, prop. n.: The US government's AI testing and standards center, part of the National Institute of Standards and Technology. It evaluates advanced AI models, American and foreign, for national-security risks and works with companies on voluntary testing., checkpointcheckpoint, n.: A saved snapshot of a model partway through or at the end of training, like a save point in a video game: training can resume from it after a crash, and finished snapshots are what people download and run., concentration of powerconcentration of power, n.: The worry that AI will put enormous economic, political or military power in the hands of a few companies, governments or individuals, beyond the reach of ordinary checks and balances.