OpenAI Set to Release Its First Open Language Model Since GPT-2

OpenAI has announced its intention to release its first open language model since GPT-2 within the coming months. This information comes from a feedback form published on the company’s website, aimed at gathering insights from developers, researchers, and members of the broader community.

Community Engagement and Feedback

The feedback form includes several questions designed to understand user preferences and experiences. Among the inquiries are:

• What features would you like in an open-weight model from OpenAI?
• What open models have you previously utilized?

OpenAI expressed enthusiasm about collaborating with the community, stating, “We’re excited to collaborate with developers, researchers, and the broader community to gather inputs and make this model as useful as possible.” Those interested in participating in feedback sessions with the OpenAI team are encouraged to indicate their interest through the form.

Upcoming Events and Demos

To facilitate feedback collection, OpenAI plans to host a series of developer events. The initial event is scheduled to take place in San Francisco in the coming weeks, followed by additional sessions in Europe and the Asia-Pacific regions. These gatherings aim to present prototypes of the upcoming model and solicit further feedback.

Responding to Competition

Amid escalating competition from other AI entities, such as the Chinese lab DeepSeek, OpenAI’s shift towards an open model is notable. These competitors are adopting transparent strategies, allowing their models to be accessible for experimentation and potential commercialization. For instance, Meta’s Llama models have achieved over a billion downloads, demonstrating the efficacy of this open approach. DeepSeek has similarly garnered a substantial global user base and attracted significant investment interest.

OpenAI’s New Vision

OpenAI’s CEO, Sam Altman, recently expressed in a Reddit Q&A that he believes the organization has not properly embraced open-source strategies in the past. He remarked, “[I personally think we need to] figure out a different open source strategy. Not everyone at OpenAI shares this view, and it’s also not our current highest priority.” Altman further indicated that while OpenAI would continue to produce advanced models, it will likely maintain a smaller lead compared to previous years.

Details on the Upcoming Model

In a social media update, Altman revealed that the forthcoming open model is expected to incorporate reasoning capabilities akin to those found in OpenAI’s o3-mini. He also assured that the model would undergo rigorous evaluation before its release, adhering to OpenAI’s preparedness framework. Additionally, he noted that extra precautions would be taken considering the model’s potential alterations by users post-release, stating, “We’re excited to see what developers build and how large companies and governments use it where they prefer to run a model themselves.”

Ensuring Transparency and Accountability

New revelations in a forthcoming book by Wall Street Journal reporter Keach Hagey suggest that Altman may have misled OpenAI executives about safety protocols concerning the model review process prior to his brief ousting in November 2023.