OpenAI, the creator of ChatGPT, has announced the launch of two new artificial intelligence models that can be customised and run locally, including on personal laptops.
Introducing GPT-oss-120b and GPT-oss-20b
The models, named GPT-oss-120b and GPT-oss-20b, belong to a category of open-weight language models. Unlike OpenAI’s flagship ChatGPT, which is built on a closed model, these new releases provide users with direct access to the model’s parameters. This allows for greater flexibility, enabling developers and organisations to tailor the AI to their specific needs.
Built for Flexibility and Local Use
OpenAI designed these models to accommodate various hardware setups. GPT-oss-120b is intended for systems equipped with a graphics card, while the more lightweight GPT-oss-20b is capable of running on laptops with just 16GB of RAM. This makes the technology more accessible to a broader range of users, from individuals to enterprise teams.
Tailored AI for Businesses and Developers
These models are particularly useful for companies that want to deploy AI on their own servers. With open access to the underlying weights, businesses can adapt the models for internal use without relying on cloud-based services. This approach offers more control over data, performance, and integration.
Capabilities and Transparency
Both GPT-oss-120b and GPT-oss-20b are capable of generating text, writing code, and performing online research when prompted. However, OpenAI has not disclosed the specific datasets used to train these models, a detail that often plays a significant role in transparency and trust within the AI community.
