To counter comparable products from US and Chinese competitors, OpenAI on Tuesday unveiled two new artificial intelligence (AI) models that are freely downloadable and customisable by users.
The ChatGPT-maker is under pressure to disclose the inner workings of its software in keeping with its non-profit roots, which is why it released the “open-weight language models” gpt-oss-120b and gpt-oss-20b.
“Since our founding in 2015, our goal has been to guarantee that artificial general intelligence (AGI) serves the interests of all people,” stated Sam Altman, CEO of OpenAI.
In the context of generative AI, an open-weight model is one that allows users to adjust the trained parameters by making them publicly available. The Chinese AI startup DeepSeek shocked the industry with its low-cost, high-performance model that boasts an open weight approach that lets users customise the technology, while Meta brags about its open-source approach to AI.
During a press briefing, OpenAI president and co-founder Greg Brockman stated, “This is the first time that we’re releasing an open-weight model in language in a long time, and it’s incredible.”
According to OpenAI, the new text-only models are easy to run on local computer systems and provide powerful performance at low cost. They are suitable for AI tasks like internet searches and computer code execution.
Altman stated, “We are really optimistic that this release will allow for new types of research and the development of new types of products.”
OpenAI said it is working with partners, including French telecommunications giant Orange and cloud-based data platform Snowflake, on real-world uses of the models. OpenAI claims that the open-weight models have been adjusted to prevent their use for malevolent intent.
Early in the year, Altman said that his organisation had been “on the wrong side of history” in terms of being transparent about the operation of its technologies. Altman also announced that OpenAI will continue to be administered as a non-profit, abandoning a contentious intention to transform into a for-profit company.
With significant investors demanding higher profits, the structural problem had turned into a source of conflict.
AI safety advocates and co-founder Elon Musk strongly opposed that strategy, suing the business he departed in 2018 on the grounds that it went against its core values.
The updated strategy will allow OpenAI’s revenue-generating division to make money while continuing to be governed by the non-profit board.
This also coincides with the launch of the two “Open” AI Reasoning Models which were released by OpenAI yesterday.
We use cookies to ensure that we give you the best experience on our website. If you continue to use this site we will assume that you are happy with it.
