OpenAI has announced that it will delay the full release of its next major AI model, GPT-5.6, after a request from the Trump administration. The decision was reportedly communicated by CEO Sam Altman during a company Q&A session on Wednesday. According to sources cited by The Information, the model will first be released in a limited preview form, granting access only to a small group of enterprise customers. During this preview period, the Trump administration itself will vet and approve each customer on a case-by-case basis.
The move highlights the growing tension between rapid AI development and government oversight. The Trump administration, which had previously signaled a “speed wins” approach to AI, has now imposed a more cautious stance. The request to OpenAI came amid fears that advanced AI models could pose national security risks, particularly if accessed by foreign entities or used for malicious purposes.
This development is part of a broader regulatory landscape that has seen the administration take increasingly firm actions against AI companies. Earlier this month, OpenAI rival Anthropic faced an ultimatum that required it to suspend access to its Mythos 5 and Fable 5 models. The government issued an export control directive prohibiting “foreign nationals” from accessing those technologies, including Anthropic's own non-US citizen employees. That heavy-handed approach alarmed many in the tech industry, as it contradicted earlier promises of fostering American AI exports.
The contrast between the treatment of OpenAI and Anthropic has not gone unnoticed. While OpenAI's delay is essentially a voluntary compliance with a government request, Anthropic received a direct ultimatum with no apparent room for negotiation. Some analysts suggest that OpenAI’s prominent position and its close ties to the current administration may have played a role in the more lenient treatment. Others point to the specific capabilities of GPT-5.6 and the nature of its intended use cases as factors that might make it less threatening than Anthropic’s models.
To understand the full significance of this delay, it is necessary to look at the history of GPT releases and the evolving regulatory environment. OpenAI had been developing GPT-5.6 for over a year, building on the success of GPT-5 and GPT-5.5. The new model promised significant improvements in reasoning, multilingual capabilities, and energy efficiency. Industry insiders had anticipated a wide rollout by mid-2026, but the federal government’s intervention has now put those plans on hold.
The Trump administration’s stance on AI has been inconsistent. On one hand, it has championed American leadership in AI, investing heavily in research and development through initiatives like the American AI Initiative. On the other hand, it has imposed restrictions on high-risk models, especially those developed by companies with international connections. The administration has also been wary of models that could be used for disinformation, cyberattacks, or autonomous warfare. The request to OpenAI appears to be part of a broader push to create a regulatory framework that allows the US to maintain its competitive edge while minimizing potential harms.
The implications of this delay extend beyond OpenAI and its customers. For enterprise clients waiting to deploy GPT-5.6, the case-by-case approval process introduces uncertainty. Companies in sectors like healthcare, finance, and defense may find their access delayed or denied depending on their use case and foreign connections. This could slow the adoption of advanced AI in critical industries and may prompt some businesses to turn to alternative providers from other countries, such as China’s Baidu or the UK’s DeepMind.
Moreover, the uneven application of regulations raises concerns about market fairness. Startups and smaller enterprises that rely on OpenAI’s API may find themselves at a disadvantage compared to larger corporations that have stronger government ties. The approval process could also create a bottleneck, as the administration may not have the capacity to handle a large volume of access requests efficiently.
On the other hand, supporters of the administration’s approach argue that careful oversight is necessary to prevent catastrophic outcomes. The potential for AI to be used in weapon systems, surveillance, or other high-stakes applications makes it imperative that the government has some control over its distribution. The case-by-case model allows for nuanced decisions based on the specific risks and benefits of each deployment.
The delay also affects OpenAI’s competitive position. With Anthropic’s models already restricted and now GPT-5.6 delayed, other AI companies like Google DeepMind and Microsoft may seize the opportunity to promote their own advanced models. However, they too may face similar scrutiny if the administration broadens its regulatory reach.
In the coming weeks, OpenAI is expected to provide more details about the preview program and how customers can apply for access. The company is also likely to engage in dialogue with government officials to refine the approval criteria. For now, the tech industry watches closely as the Trump administration’s AI policies take shape, with the hope that regulations will eventually become more predictable and uniform.
Source:The Verge News