OpenAI Joins the Open-Source AI Movement, Releasing Free and Customizable Models

OpenAI takes on Meta by launching free and customisable AI models

Developer of ChatGPT says the tools will be 'for wide benefit', echoing announcement by Mark Zuckerberg OpenAI is taking on Mark Zuckerberg's Meta by launching its own freely available artificial intelligence models. The ChatGPT developer has announced two open-weight large language models, which are free to download and can be customised by a user. Meta's Llama models are available on a similar basis, and OpenAI's move marks a departure from ChatGPT, which is based on a "closed" model that cannot be customised. Sam Altman, OpenAI's chief executive, said the company was excited to add to a stack of freely available AI models "based on democratic values ... and for wide benefit". He added: "We're excited to make this model, the result of billions of dollars of research, available to the world to get AI into the hands of the most people possible." Zuckerberg, similarly, has said that making models freely available and customisable will "ensure that more people around the world have access to the benefits and opportunities of AI" and that power is not concentrated in the hands of a small number of companies. However, Meta has said his company may need to be "careful" about taking a similar approach with highly advanced AI models. DeepSeek, a Chinese rival to OpenAI and Meta, has also released powerful models that can be downloaded and customised freely. OpenAI said the two models, called gpt-oss-120b and gpt-oss-20b-two, outperform similarly sized models on reasoning tasks, with the larger, 120b model achieving a near-equal performance to its o4-mini model in terms of core reasoning. Open AI added that during the testing process it had created "maliciously fine-tuned" versions of the models to simulate biological and cybersecurity threats but found that they were "unable to reach high capability levels". The existence of freely available, powerful AI models that can be customised by developers has alarmed some experts who have warned that they could be adapted to cause serious harm, such as developing bioweapons. Meta has defined its Llama models as "open source", meaning they can also be downloaded and customised freely, while their training datasets, architecture and training code are also provided. However, the Open Source Initiative, a US-based industry organisation, has said the restrictions Meta places on using its models mean it does not qualify as fully open source. OpenAI has used the term "open weight", effectively a notch down from open source, which means the model can still be fine tuned - but its workings are not fully transparent to the developer. OpenAI's announcement came amid speculation that a new version of the model that underpins ChatGPT could be released soon. On Sunday Altman shared a screenshot of what appeared to be the company's latest AI model, GPT-5. Meanwhile, Google has outlined its latest step towards artificial general intelligence (AGI) with a new model that allows AI systems to interact with a convincing simulation of the real world. The Genie 3 "world model" could be used to train robots and autonomous vehicles as they engage with realistic recreations of environments such as warehouses, according to Google. The US technology company's AI division, Google DeepMind, argues that world models are a key step to achieving AGI, a hypothetical level of AI where a system can carry out most tasks on a par with humans - rather than just individual tasks such as playing chess or translating languages - and potentially do someone's job. DeepMind said such models would play an important role in the development of AI agents, or systems that carry out tasks autonomously. "We expect this technology to play a critical role as we push toward AGI, and agents play a greater role in the world," DeepMind said.

OpenAI to give away some of the technology that powers ChatGPT

SAN FRANCISCO -- In a move that will be met with both applause and hand-wringing from artificial intelligence experts, OpenAI said Tuesday that it was freely sharing two of its AI models used to power online chatbots. Since OpenAI unveiled ChatGPT three years ago, sparking the AI boom, it has mostly kept its technology under wraps. But many other companies, looking to undercut OpenAI, have aggressively shared their technology through a process called open source. Now, OpenAI hopes to level the playing field and ensure that businesses and other software developers stick with its technology. OpenAI's shift adds more fuel to a long-running debate between researchers who believe it is in every company's interest to open-source their technology, and national security hawks and AI safety pessimists who believe American companies should not be sharing their technology. The China hawks and AI worriers appear to be losing ground. In a notable reversal, the Trump administration recently allowed Nvidia, the world's leading maker of the computer chips used to create AI systems, to sell a version of its chips in China. Many of the San Francisco company's biggest rivals, particularly Meta and Chinese startup DeepSeek, have already embraced open source, setting OpenAI up as one of the few AI companies not sharing what it was working on. The models being offered by OpenAI, called gpt-oss-120b and gpt-oss-20b, do not match the performance of OpenAI's most powerful AI technologies. But they still rank among the world's leading models, according to benchmark test results shared by the company. If people use those newly open-source models, OpenAI hopes they will also pay for its more powerful products. "If we are providing a model, people are using us," Greg Brockman, OpenAI's president and one of its founders, said in an interview with The New York Times. "They are dependent on us providing the next breakthrough. They are providing us with feedback and data and what it takes for us to improve that model. It helps us make further progress." Open source has been a common practice among software companies for decades. As OpenAI and other companies began developing the kind of technology that would eventually drive chatbots like ChatGPT nearly a decade ago, they often open-sourced them. "If you lead in open source, it means you will soon lead in AI," said Clément Delangue, chief executive of Hugging Face, a company that hosts many of the world's open-source AI projects. "It accelerates progress." But after OpenAI shared a technology called GPT-2 in late 2019, it stopped open-sourcing its most powerful systems, citing safety concerns. Many of OpenAI's rivals followed its lead. When OpenAI unveiled ChatGPT in late 2022, a growing chorus of AI experts argued that open-source technologies could cause serious harm. This kind of technology can help spread disinformation, hate speech and other toxic language. Many researchers also worry that they could one day help people build bioweapons or wreak havoc as governments and businesses connected them to power grids, stock markets and weapons. But the public conversation started to shift in 2023, when Meta shared an AI system called LLama. Meta's decision to go against the grain fueled a growing open-source ecosystem in the United States and other parts of the world. By late 2024, when DeepSeek released a technology called V3, China had shown that its open-source systems could challenge many of the leading U.S. systems. OpenAI said that it has released open-source systems in part because some businesses and individuals prefer to run these kinds of technologies on their own computer hardware, rather than over the internet. One of the new systems, gpt-oss-20b, is designed to run a laptop. The other, gpt-oss-120b, requires a more powerful system equipped with the specialized computer chips used to build the leading AI systems. Brockman acknowledged that AI can be used to both harm and empower people. But he said that the same is true of any powerful technology. He said that OpenAI has spent months -- even years -- building and testing its new open-source systems in an effort to reduce any harm. The debate over open source is expected to continue as companies and regulators weigh the potential harm against the power of the time-tested tech method. Many companies have changed strategy over the years and will continue to do so. After creating a new superintelligence lab, Mark Zuckerberg and his fellow executives at Meta are considering their own shift in strategy. They might abandon the AI technology that the company has freely shared with researchers and businesses, called Behemoth, and move toward a more guarded strategy involving closed-source software.