OpenAI to Release 'Very, Very Good' Open-Weight Language Model
2 minutes
OpenAI has announced plans to release a "powerful new open-weight language model with reasoning" in the coming months, marking their first such release since GPT-2. CEO Sam Altman emphasized the company's excitement about making this "a very, very good model" while highlighting the importance of developer input in shaping its final form.
A Long-Awaited Return to Open Weights
While OpenAI has been considering this release "for a long time," other priorities had taken precedence until now. The timing feels particularly important, according to Altman, as the organization returns to making AI models fully accessible to researchers and developers.
An open-weight model allows developers to access and modify its internal components, unlike OpenAI's current models like GPT-4, which are only available through APIs. This access means organizations can examine how the model works internally, modify it for their specific needs, and conduct thorough security audits before deployment. Most importantly, researchers and developers can build new applications directly on top of the model's architecture, fostering innovation in ways that aren't possible with closed systems.
Responsible Development and Release
OpenAI is taking a measured approach to the release. The model will undergo evaluation through their preparedness framework, with additional considerations given to its open-weight nature and potential post-release modifications. The organization recognizes that this model will be modified after release and is doing extra work to ensure it's ready for such adaptations.
Before finalizing the release, OpenAI is actively seeking developer feedback to make the model "maximally useful." The company has organized a series of developer events starting in San Francisco, followed by sessions in Europe and Asia-Pacific regions. These events will give developers the opportunity to:
- Provide early feedback on the model's design
- Test early prototypes
- Help shape the model's final capabilities
Looking Ahead
OpenAI is particularly excited to see what developers will build with this model. The company anticipates significant interest from large organizations and governments that prefer to run models internally.
Valeriia Kuka
Valeriia Kuka, Head of Content at Learn Prompting, is passionate about making AI and ML accessible. Valeriia previously grew a 60K+ follower AI-focused social media account, earning reposts from Stanford NLP, Amazon Research, Hugging Face, and AI researchers. She has also worked with AI/ML newsletters and global communities with 100K+ members and authored clear and concise explainers and historical articles.