June 11 News.OpenAI ofOpen Source ModelThere's been a new tweak to the release program. The company's CEO, Sam Altman (Sam Altman) announced in a Tuesday post on the X platform that the open-source model, originally planned for early summer this year, is expected to be delayed until late summer and will not be available to the public in June.
1AI notes that in his post, Altman writes, "We're going to take a little more time for the open source weighting model, which means that people can expect it late this summer, but not in June." He also mentioned thatOpenAI's research team has made some "unexpected and quite surprising" discoveries during the development process that will make the final open source model "well worth the wait," but it will take more time to refine.
OpenAI's open-source models were originally planned to have similar "reasoning" capabilities as the company's O-series models, and were intended to outperform other existing open-source reasoning models, such as DeepSeek's R1 model. Since OpenAI first announced that it would be releasing an open source model, competition in the space has intensified. Just Tuesday, Mistral, another AI lab that regularly publishes open-source models, launched Magistral, its first line of AI inference models, and in April, Qwen, a Chinese AI lab, released a set of hybrid AI inference models that switch between in-depth reasoning about a problem and giving a quick, traditional answer.
In an effort to score better on performance benchmarks, OpenAI is also considering adding some sophisticated features to its open-source models to boost their competitiveness. As TechCrunch previously reported, OpenAI's leadership has discussed enabling the open-source model to connect to the company's cloud-hosted AI models to handle complex query requests. However, it's unclear whether these features will end up in the open source models.
Altman has previously said that OpenAI is on the "wrong side of history" when it comes to open source models. To correct this image, the company is under intense pressure to deliver an open source model that can compete with the industry's top open source products.