OpenAI plans to release open-weight language model in coming months

OpenAI intends to launch its initial open-weight language model featuring reasoning abilities since GPT‑2 in the near future, CEO Sam Altman stated on Monday.

The trained parameters or weights of an open-weight language model are available to the public, allowing developers to examine and optimize the model for particular tasks without needing the initial training data.

Altman mentioned in his post on X that the company will discuss with developers ways to enhance the utility of an open-weight language model.

The initial event is scheduled for San Francisco in a few weeks, with subsequent sessions in Europe and the Asia-Pacific regions. In February, Altman stated that the company would streamline its AI offerings while he presented a new strategy for the latest models.

A source familiar with the situation stated on Friday that OpenAI, which is supported by Microsoft (MSFT.O), must convert into a for-profit organization by year-end to obtain the complete $40 billion investment spearheaded by SoftBank Group (9984.T).

The startup has stated that shifting to a for-profit organization is necessary to obtain the funding required for creating the top AI models.

In October, OpenAI completed a funding round worth $6.6 billion.