OpenAI delays release of open-source AI model, promises ‘amazing’ upgrade later this summer

OpenAI has announced a delay in the release of its long-anticipated open-weight AI model, according to CEO Sam Altman. Initially scheduled for June, the debut has now been shifted to later this summer. Altman emphasized that the extra time is necessary to enhance the model's performance, which he describes as a substantial improvement. In a post on X, Altman stated, 'We are going to take a little more time with our open-weights model, i.e. expect it later this summer but not June.' He went on to reveal that the research team has achieved something unexpected and extraordinary, suggesting that the wait will be worthwhile. This postponement occurs at a time when competitors are rapidly releasing their own open-source AI models, increasing the pressure on OpenAI to introduce a product that excels in both transparency and capability. The forthcoming model is anticipated to incorporate reasoning abilities akin to those found in OpenAI's proprietary series, aiming to surpass rivals like DeepSeek’s R1. Recently, French startup Mistral introduced a series of reasoning-focused models called Magistral. Simultaneously, the Chinese lab Qwen launched hybrid models capable of switching between fast responses and more methodical 'chain-of-thought' processing, setting a new standard for advanced AI functionality. While OpenAI hints at integrating advanced features into its open model, such as connectivity with its cloud-based systems for tackling complex queries, it remains uncertain whether these enhancements will be part of the inaugural release. The delay also raises questions about OpenAI's reputation, as Altman acknowledged the criticism the company has received for moving away from open-sourcing their leading models. There is growing concern among researchers and developers that transparency is critical for ensuring safety, reproducibility, and fostering innovation in the field of AI.