OpenAI CEO Sam Altman said that the company will "take a little more time" with its first open-weight language model since GPT-2.
"Expect it later this summer but not June," Altman wrote on X, saying that the delay is due to the research team having done "something unexpected and quite amazing." He added that the company thinks the wait will be "very very much" worth it.
Open-weight language models have their trained parameters (weights) publicly available, meaning that developers can download, customize and use these models for specific tasks without needing original training data. Open-source models, on the other hand, provide direct access to the complete source code, dataset and methodology.
