.jpeg?width=1200&auto=webp&trim=679%2C494%2C384%2C1099)
OpenAI has launched a new artificial intelligence model that requires no internet connection or massive data centres to operate.
The ChatGPT creator said the latest gpt-oss-120b and gpt-oss-20b models will instead run locally on consumer devices like laptops or smartphones.
Using fewer compute resources means the new models have a far lower environmental impact than conventional cloud-based models, however one of their limitations is that they provide text-only answers.
OpenAI boss Sam Altman described gpt-oss as a “big deal” that is the result of billions of dollars of research.
“We believe this is the best and most usable open model in the world,” he wrote on X. “We’re excited to make this model... available to the world to get AI into the hands of the most people possible.
“We believe in individual empowerment. Although we believe most people will want to use a convenient service like ChatGPT, people should be able to directly control and modify their own AI when they need to, and the privacy benefits are obvious.”
Mr Altman claimed that the new models would enable new kinds of research, as well as lead to new inventions.
“We expect a meaningful uptick in the rate of innovation in our field, and for many more people to do important work than were able to before,” he said.
It is the first time since GPT-2 – a precursor to ChatGPT – that OpenAI has launched an open-weight model that runs locally on a device.
Mr Altman cited safety concerns for not releasing this type of model, despite them being lower cost, more customisable, and better for the environment.
OpenAI tested the new models with external safety firms in order to minimize risks in areas like biological weapons and cyber security.
Both of the new models are available on developer platforms like Hugging Face and Databricks as a free download from Wednesday.
