OpenAI delays the release of its open model, again
OpenAi CEO Sam Altman said on Friday that the company postponed the release of its open model, which was pushed back a month earlier in this summer. OpenAi was planning to release the model next week, but Altman said that the company pushes it back for further safety tests for an indefinite period of time.
“We need time to perform extra safety tests and to assess areas with a high risk. We are not sure how long it will take us,” Altman said in a Post on X. “Although we trust that the community will build great things with this model, as soon as the weights are out, they cannot be withdrawn. This is new to us and we want to do well.”
The Open Model release of OpenAi is one of the most expected AI events of the summer, in addition to the expected release of the chatgpt maker of GPT-5. In contrast to GPT-5, the OpenAI’s open model will be available for developers to freely download and perform locally. Both launches will try to show that it is still the leading AI lab of Silicon Valley – an increasingly difficult task such as Xai, Google DeepMind and anthropic investing billions of dollars in their own efforts.
The delay means that developers have to wait a little longer to try that the first open model that OpenAi has released in years. WAN reported earlier that the Open Model of OpenAI is expected to have similar reasoning options if the O-series models of the company, and that OpenAi was planning to be the best compared to other open models.
The ecosystem of Open AI models became a little more competitive this week. Earlier on Friday, Chinese AI Startup Moonshot Ai launched Kimi K2, An open AI model of a trillion parameter that performs better than the GPT-4.1 AI model from OpenAI on various agent coding benchmarks.
In June, when Altman announced the first delays around the Open Model of OpenAi, he noticed that the company had achieved something “unexpected and rather amazing”, but did not respond to what that was.
“In terms of capacities, we think that the model is phenomenal – but our bar for an open source model is high and we think we need a little more time to ensure that we release a model that we are proud of during each axis,” said Aidan Clark, OpenAi’s VP of research that the Open Modellam team is in one Post on X Friday.
TechCrunch reported earlier that OpenAi leaders have discussed to enable the open AI model to connect to the company’s cloud-hosted AI models for complex questions. However, it is unclear whether these functions will reach the last open model.
