Sam Altman launches GPT-oss, OpenAI's first open-weight AI language model in over 5 years

In a significant move towards transparency in artificial intelligence, OpenAI has introduced GPT-oss, a groundbreaking family of language models featuring open weights. Announced by CEO Sam Altman, this initiative allows the models to operate on high-end laptops and smartphones, marking a shift towards more accessible AI technology. The concept of 'open weights' means that the fully trained parameters of these models are available for public download. This allows users to run, inspect, and fine-tune the AI on their own devices. Altman emphasized the potential of GPT-oss, describing it as 'the best and most usable open model in the world' in a post on X. The GPT-oss family includes two distinct models: gpt-oss-120b and gpt-oss-20b. The latter is tailored for operation on most desktops and laptops, while the former is optimized for high-performance computing environments. Altman noted that GPT-oss delivers impressive real-world performance, rivaling that of the recently launched o4-mini. Interestingly, this announcement comes shortly after competitor Anthropic unveiled the Claude Opus 4.1, although GPT-oss is not the much-anticipated ChatGPT-5, which could debut later this week. This release marks OpenAI's first foray into open-weight models since the launch of GPT-2 in 2019. Altman expressed optimism about the future implications of this release, stating that it could foster innovative research and lead to the development of new products. He believes this will accelerate the pace of innovation in the AI sector, enabling a broader range of individuals to contribute meaningfully to the field. OpenAI has previously hinted at its commitment to developing powerful open-source models, aiming to surpass existing offerings from competitors like Meta, which has made its LLAMA AI models available with open weights.