OpenAI introduces GPT-OSS, its first open-weight model in over six years, available in two variants: 120-billion and 20-billion parameters. The larger model can operate on a single Nvidia GPU, while the smaller variant needs only 16GB of memory. This model promotes extensive customization and is released under the Apache 2.0 license across several platforms, including Hugging Face and Azure. CEO Sam Altman previously cited safety as a reason for delayed releases but now acknowledges developer demand for open models, signaling a strategic shift in OpenAI's approach to AI accessibility.
OpenAI is releasing a new open-weight model dubbed GPT-OSS that can be downloaded for free, be customized, and even run on a laptop.
The model comes in two variants: 120-billion-parameter and 20-billion-parameter versions.
Both model versions are being released today via platforms like Hugging Face, Databricks, Azure, and AWS under the Apache 2.0 license.
This is OpenAI's first open-weight model in over six years, years before the debut of ChatGPT.
Collection
[
|
...
]