OpenAI's open-source AI model has been postponed indefinitely once again. Safety risks and a review of high-risk applications have been cited as reasons, leading the company to avoid rushing the release.
OpenAI: OpenAI, a leader in the field of artificial intelligence, has once again delayed the much-anticipated release of its open-source AI model. The model was initially scheduled to launch in June 2025, but now company CEO Sam Altman himself has confirmed that the release has been postponed indefinitely. The reasons cited behind this decision are security testing and a review of high-risk areas. This news is spreading rapidly in the tech world, raising questions about the level of precautions OpenAI is taking and why.
Why is the launch date being repeatedly postponed?
This is not the first time the model's release has been delayed. It was originally planned for June 2025, but even then, it was pushed back. Now, in July, another indefinite postponement has occurred. This model was considered OpenAI's first open-source AI project since 2019, promising new possibilities for developers.
Security becomes the top priority
Altman posted on the social media platform X (formerly Twitter): 'We were planning to launch our open-weight AI model next week. However, we need more time to conduct further security testing and review high-risk applications. We also don't know how long this might take.' This statement makes it clear that OpenAI does not want to rush this time. In an era where the misuse of AI has become easier, releasing such powerful technology openly could pose a significant risk.
Open-Weight Model: What is it?
The model OpenAI was planning to release is an 'open-weights' model. This means that:
- Developers and tech enthusiasts can download this model.
- They can customize its parameters and weights.
- They can run it on their local systems.
However, it should be noted that the model's training methodology, data, and architecture will not be made public. This is to ensure that no one can copy the model exactly and use it in an unsafe or unethical manner.
Increasing competitive pressure
OpenAI's delay comes at a time when its competitors such as:
- Google (Gemini)
- Anthropic (Claude)
- Meta (LLaMA)
- Alibaba (ERNIE
Alibaba has recently joined the race by open-sourcing its ERNIE model. In this situation, this decision by OpenAI could also be a strategic move—allowing them to return to the market with a safer and more stable release.
Open-source versus Proprietary Models: A Dual Strategy
OpenAI's step points towards a dual strategy:
- Premium models (like GPT-5) will be reserved for corporate and commercial use only.
- While the open-source model will attract developers and the technical community.
With this balance, the company wants to keep both groups—consumers and developers—connected. However, it is not yet clear which model will be released first and when.
What did the open model feedback reveal?
In April 2025, OpenAI released an 'Open Model Feedback Form' to understand what developers expect from the model. Based on this feedback, it was believed that:
- Developers want the freedom to fine-tune.
- Some users want to use it for academic and research purposes.
- Many people also had concerns about security and responsible AI development.
- That's why OpenAI is not rushing now.
