infotech

OpenAI Delays The Launch Of Its Open Model Again

OpenAI CEO Sam Altman announced on Friday (July 11th) that the company has again postponed the launch of its open model, which had been delayed since earlier this summer. 

As reported by TechCrunch on Friday (July 11th), OpenAI originally planned to release the model next week, but Altman stated that the launch has been postponed indefinitely for further security testing. 

“We need time to conduct more security testing and review high-risk areas. We don't know exactly how long that will take,” Altman wrote on the X platform. 

“We believe the community will create amazing things with this model, but once it's released, we can't take it back. This is new ground for us, and we want to get it right,” he continued. 

The launch of OpenAI's open model is one of the most anticipated AI moments this summer, along with the equally anticipated launch of GPT-5 from the creators of ChatGPT. 

Unlike GPT-5, OpenAI's open model will be available for developers to download and run locally. With these two launches, OpenAI is attempting to demonstrate that it remains Silicon Valley's leading AI lab — a challenge made even more challenging by the significant investments made by xAI, Google DeepMind, and Anthropic in its own development. 

The delay means developers will have to wait longer to try out the first open model released by OpenAI in recent years. 

TechCrunch previously reported that this open model is expected to have reasoning capabilities comparable to OpenAI's "o" series of models, and is designed to be the best among other open models. 

The open AI model ecosystem has become increasingly competitive this week. Earlier Friday, Chinese AI startup Moonshot AI launched Kimi K2, an open AI model with one trillion parameters that outperformed OpenAI's GPT-4.1 model in several agentic coding benchmarks. 

Last June, when Altman announced the initial delay of this open model, he said the company had achieved something "unexpected and quite amazing," though he did not elaborate on what he meant. 

“In terms of capabilities, we think this model is fantastic — but our standards for open source models are very high, and we felt we needed more time to ensure it was worthy of praise in every aspect,” said Aidan Clark, OpenAI’s VP of Research who leads the open source model development team, in a post on X on Friday. 

TechCrunch also previously reported that OpenAI leaders were considering connecting this open source AI model with the company’s cloud-hosted AI models to handle complex questions. However, it is unclear whether that feature will be included in the final version of the open source model. 

---

Tag: OpenAI TechCrunch X Platform