OpenAI to release its first ‘open’ language model since GPT-2 in 2019

Key Points

  • OpenAI plans to release an "open-weight" language model, allowing developers to use, modify, and deploy it on their own hardware.
  • This model will be the first "open" release since GPT-2 in 2019, marking a shift from their recent closed systems like GPT-3 and GPT-4.
  • OpenAI CEO Sam Altman announced the upcoming release and mentioned plans for developer events to gather feedback and showcase prototypes.
  • The model aims to cater to large companies and governments who prefer running models on their own infrastructure.

Summary

OpenAI is gearing up to release its first "open-weight" language model since the partial release of GPT-2 in 2019, marking a significant shift from their recent closed systems. This new model will be available for anyone to use, download, modify, or deploy, providing developers with the flexibility to run it on their own hardware. CEO Sam Altman announced this move on X, emphasizing the importance of gathering feedback to maximize the model's utility. The decision to release an open model comes after prioritizing other projects, but now feels crucial. Developer events are planned to start in San Francisco, with subsequent sessions in Europe and the Asia-Pacific region. This release is part of a broader trend in the AI industry, with competitors like Alibaba, Google, and Meta also advancing their AI models, highlighting the intensifying AI arms race.

cointelegraph
April 2, 2025
Stocks
Read article

Related news