
OpenAI has announced plans to introduce its latest open-weight language model, marking the first release of its kind since the GPT-2 model. The CEO of the AI research organisation Sam Altman shared the update via a post on social media platform X.
An open-weight model differs from open-source models, which provide complete access to the source code, training data, and development processes. This upcoming model will allow developers to access and modify the model’s trained parameters, facilitating the customisation of the AI for specific tasks without the original training dataset.
Altman noted that OpenAI is initiating conversations with developers to determine the most beneficial applications of the new model. The first of these discussions is scheduled to take place in San Francisco, with subsequent meetings planned for Europe and the Asia-Pacific region.
“Before release, we will evaluate this model according out our preparedness framework, like we would for any other model and we will do extra work given that we know this model will be modified post-release,” Altman said in the post.
OpenAI valuation reaches $300bn
Meanwhile, OpenAI has secured a significant $40bn investment at a $300bn post-money valuation, with Japan’s SoftBank leading the funding round. The firm plans to use the new funding to further AI research, enhance its computing infrastructure, and improve tools for the platform’s user base, which includes 500 million people interacting with ChatGPT weekly.
SoftBank’s investment includes a $10bn initial contribution in mid-April and a conditional $30bn in December, dependent on OpenAI’s transition to a for-profit organisation by the end of the year.
“We’re excited to be working in partnership with SoftBank Group—few companies understand how to scale transformative technology like they do,” OpenAI said in a statement. “Their support will help us continue building AI systems that drive scientific discovery, enable personalised education, enhance human creativity, and pave the way toward AGI that benefits all of humanity.”
Additionally, OpenAI has indicated that its advanced AI agent, ‘Deep Research’, currently available only to Plus and Enterprise customers, will soon be accessible to all users, including those who do not subscribe to ChatGPT services, according to a BleepingComputer report. Deep Research is designed to perform extensive, multi-step internet research tasks more rapidly than traditional human-led efforts.
Earlier this year, OpenAI introduced ChatGPT Gov, a specialised version of its AI platform, designed for US government agencies.