WASHINGTON — OpenAI has halted the rollout of its newest AI model, GPT‑6.1 Astra, citing safety concerns, the company said on Tuesday. The decision follows growing scrutiny of AI systems that can generate realistic text and images. The move marks a rare retreat by a leading AI firm amid regulatory pressure.
Model details and safety concerns
GPT‑6.1 Astra was expected to build on the capabilities of GPT‑5, offering advanced reasoning and multimodal features. OpenAI had planned to release the model to a limited set of partners before a wider public launch. The company said an internal safety review identified risks related to hallucinations, disinformation potential, and misuse in high‑stakes domains.
OpenAI’s safety review included input from external experts and a review of the model’s alignment with the company’s policy framework. The review concluded that the model’s safety mitigations were insufficient to prevent certain types of harmful outputs, prompting the pause.
Regulatory backdrop
The decision comes as regulators in the United States, the European Union, and other jurisdictions tighten rules on AI. The U.S. Federal Trade Commission has issued guidance on deceptive AI content, while the EU’s AI Act imposes strict requirements on high‑risk systems. OpenAI has previously faced scrutiny over GPT‑4’s content moderation and the company has emphasized its commitment to safety in public statements.
Impact on OpenAI’s roadmap
OpenAI said it will continue to work on GPT‑6.1 Astra but will delay its release until safety concerns are fully addressed. The company’s CEO, Sam Altman, stated that the organization remains focused on developing safe AI and that the pause is part of a broader strategy to ensure responsible deployment.
The pause does not affect existing OpenAI products, including ChatGPT and the API services that rely on GPT‑4. OpenAI’s revenue from these services is expected to remain stable in the short term, although the company has not released updated financial projections.
Industry reaction
Other AI firms have also paused or delayed releases in response to safety and regulatory concerns. Anthropic, Google, and Microsoft have all announced increased safety testing for upcoming models. The trend reflects a growing emphasis on responsible AI development across the industry.
OpenAI’s decision may influence competitors’ timelines and could prompt further regulatory scrutiny. The company’s pause is a signal that safety concerns are taking precedence over rapid product rollout.
