Sam Altman Announces OpenAI's New Open-Weight AI Model

On Monday, OpenAI CEO Sam Altman revealed plans to launch a new open-weight language model equipped with advanced reasoning capabilities in the coming months. This initiative seems influenced by the notable success of DeepSeek’s R1 model and the increasing recognition of Meta Platforms, Inc. and its Llama models.

What Happened: Altman shared the announcement via a post on X, formerly known as Twitter. He stated that this forthcoming model would be the first of its type since GPT-2, featuring publicly accessible trained parameters or weights that allow developers to customize the model for various tasks.

In his post, Altman noted, “We've been thinking about this for a long time, but other priorities took precedence. Now it feels important to do.”

He also mentioned that OpenAI intends to gather feedback from developers through a series of global events. The initial event is scheduled to occur in San Francisco, followed by additional sessions in Europe and the Asia-Pacific region.

“We still have some decisions to make, so we are hosting developer events to gather feedback and later play with early prototypes,” Altman added, expressing anticipation for the new model.

Subscribe to the latest tech developments to stay informed.

OpenAI currently provides access to its AI primarily through a chatbot and cloud services. In contrast, models like R1, Llama, and other open-weight alternatives can be freely downloaded and adapted. The weights of a model symbolize the internal values of a neural network, which are determined during its training process. Open-weight models represent a budget-friendly option that can be tailored for specific needs, including processing sensitive or confidential information.

Since the launch of DeepSeek’s model in January, Altman has indicated that OpenAI might have been “on the wrong side of history” regarding open models, suggesting a potential shift in its approach.

OpenAI appears poised to showcase its capacity to train the new model cost-effectively, particularly considering DeepSeek's ability to develop its model at a significantly lower expense than most expansive AI models available today.

This news has been well-received in the industry. Clement Delangue, co-founder and CEO of Hugging Face, a company known for supporting open AI models, emphasized, “This is amazing news. With DeepSeek, everyone's realizing the power of open weights.”

In addition to these developments, OpenAI is transitioning to a for-profit model by the end of the year to facilitate a $40 billion funding round led by SoftBank Group, which aims to back future AI model development.

This strategy follows the close of a $6.6 billion funding round in October.

Meanwhile, Meta has experienced substantial success with its family of AI models, termed ‘Llama,’ which recently reached a milestone of 1 billion downloads. This achievement was highlighted by Meta CEO Mark Zuckerberg, who emphasized the increasing trust in open-source AI models.

OpenAI, technology, AI, DeepSeek, Meta