OpenAI launches reasoning LLM that you can download and tweak
Briefly

OpenAI launched gpt-oss, a new large language model that is open-weight, allowing researchers to download and customize it. This model has competitive performance on certain tasks, available in two sizes suitable for local and offline use. This capability enhances the potential for working with sensitive data in private networks. The launch intensifies competition among open-source models, especially against emerging Chinese counterparts which may already outperform US models in some metrics. The development of gpt-oss was not a direct reaction to the success of these models, as stated by company representatives.
"The firm, based in San Francisco, California, detailed the system in a blog post and a technical description on 5 August. On some tasks, gpt-oss performs almost as well as the firm's most advanced models."
"The release of gpt-oss comes at a time when powerful open-weight models from Chinese firms, such as Hangzhou-based DeepSeek and Beijing-based Moonshot AI, are gaining traction among researchers."
Read at Nature
[
|
]