Open announced that the release of its open source artificial intelligence (AI) model was announced. On Saturday, the company’s CEO, Sam Altman, said that due to the extra time needed to run the security test, the release of the AI model was delayed indefinitely. This is the second delay after the company postponed the release in June. It is said that the first open source release of San Francisco -based AI firm has been released for the first open source release, and the first time the AI model based on reasoning. In April, the company shared an “open model feedback” form to understand what to understand what the developer wants to do with the model.
Delayed due to the release of Openi’s Open Source AI model
In a post on X (previously known as Twitter), Altman said, “We are planning to launch our openweight model next week. We are delaying it. We need time to run additional security tests and review the high -risk areas. We are not sure we will be sure yet.”
The announcement comes at a time when Openi is also ready for the next generation of its Frontier AI model, GPT5. In recent months, its rivals, such as Google, Anthropic, and Alibaba, have released the flagship grade AI models, which have done well on several benchmarks on Open AI’s latest models.
In addition, these companies (besides Meta and Z) have also made significant contributions to the open community. Alibaba was the latest to join the trend as he opened the company’s AII models.
Openi’s planned open source contribution was likely a strategic move that was a leader in the position as a leader in the position between the two businesses and users (with the GPT -5 proprietary model), and the developers (with the open source model). Nor has it arrived so far.
According to a Tech Crunch report, the AI firm is planning to issue an AI model compared to his O -series larger language model (LLM). However, the company is only planning to release the weight of the model to the open community.
This means that developers and fans will be able to download the model, fix its parameters and run it locally. However, they will not learn about its training methods and architecture to develop a copy of similar models, or build it above.
