OpenAI, the company behind ChatGPT, has announced plans to release a new open-weight language model with reasoning capabilities in the coming months.
CEO Sam Altman shared the update on social media, noting that this will be the company’s first open-weight model since GPT-2. Alongside the announcement, OpenAI launched a feedback form, inviting developers, researchers, and the broader tech community to help shape the model before its official release.
OpenAI is seeking input through developer feedback sessions to improve the model’s functionality and safety. The company has scheduled several events, starting in San Francisco, followed by sessions in Europe and Asia-Pacific, where participants will get early access to prototypes. These gatherings aim to collect real-world insights from users and developers, especially those who prefer deploying AI systems independently.
The model will undergo safety checks using OpenAI’s preparedness framework, which is part of the company’s standard process. Sam Altman stated that the model may be adjusted after launch based on public feedback and real-world testing.
Why Open-Weight Matters
Unlike closed models where only the company controls usage, open-weight models allow organizations, developers, and researchers to run the system on their own hardware. This approach appeals to institutions like large companies or governments that need more control over how AI is used. It also supports greater transparency and fosters innovation within the open-source community.
$40 Billion Funding to Fuel Growth
The announcement comes shortly after OpenAI raised $40 billion in funding, reaching a $300 billion post-money valuation, making it the largest private tech deal to date. SoftBank led the round, with support from major investors like Microsoft, Thrive, Altimeter, and Coatue.
According to OpenAI, a significant portion of the funding — around $18 billion — will be used for the Stargate infrastructure project, which aims to build a nationwide network of AI-focused data centers. The rest will help accelerate AI research and improve the computing power needed to train and run new models.