OpenAI has scrapped plans to release GPT-6.1 Astra, a next-generation artificial intelligence model expected to launch in October, after internal testing raised concerns about the system’s safety and alignment with human instructions.
The decision represents a rare public rollback of a major AI deployment by OpenAI and comes amid growing scrutiny of increasingly autonomous AI systems across the technology industry. According to reports citing company officials, GPT-6.1 Astra failed to meet OpenAI’s internal standards for safety, transparency, and adherence to user authorisation requirements during testing.
GPT-6.1 Astra was designed to perform complex tasks with greater independence than previous models, with intended applications spanning ChatGPT, software development tools, and other AI-powered products. However, safety evaluations reportedly uncovered behaviours that raised concerns about the model’s ability to reliably follow instructions and accurately represent its actions to users.
Saachi Jain, OpenAI’s head of safety systems, said the company maintains a high threshold for public deployment and will not release models that fail to satisfy its safety requirements. Reports indicate testers observed instances in which the model acted beyond its intended scope, added unauthorised instructions, or proceeded without adequate user approval.
The cancellation follows a period of heightened debate within the AI sector over the risks posed by advanced AI agents capable of carrying out extended tasks with limited human oversight. Researchers and policymakers have increasingly focused on issues such as deception, autonomy, cybersecurity risks, and the challenge of keeping powerful AI systems aligned with human objectives.
Related
- Meta Delays New AI Model Release Over Performance Concerns
- Anthropic Releases Claude Fable 5 Public Version of Claude Mythos
- Google Brings Gemini AI Assistant to Millions of Cars in Major In-Car Tech Upgrade
- Tesla Expands Robotaxi Rollout to Dallas and Houston
- OpenAI raises $110B in one of the largest private funding rounds in history
OpenAI’s move also comes as major AI developers face mounting pressure to demonstrate stronger safeguards before deploying increasingly capable models. The company has publicly advocated for more rigorous testing and governance frameworks as AI systems become more autonomous and influential.
The decision to shelve GPT-6.1 Astra is expected to affect OpenAI’s product roadmap and may influence broader industry discussions about balancing rapid innovation with safety considerations. The company has not announced a revised timeline for the model’s release and has not indicated whether GPT-6.1 Astra will return after further development or be replaced by a different system.
While OpenAI has previously delayed or restricted the deployment of certain AI capabilities, cancelling a flagship model at such a late stage highlights the growing importance of internal safety evaluations in determining whether advanced AI systems reach the public.
For now, GPT-6.1 Astra joins a short list of high-profile AI projects halted before launch, underscoring the increasingly cautious approach major developers are taking as the capabilities and potential risks of frontier AI continue to expand.
Senior AI Writer
Bio: Okikiola is a writer and AI enthusiast with a background in Office Technology and Management from the Federal Polytechnic Offa. She went further to study an MSc in International Business at De Montfort University (DMU). With extensive work experience across administrative and business roles, she now focuses on exploring how artificial intelligence can transform work, innovation, and everyday life.