OpenAI CEO Sam Altman has announced three new artificial intelligence models named Sol, Terra, and Luna, but they won't reach the public immediately due to U.S. government oversight of their release. This represents a significant departure from how AI companies have traditionally launched products, signaling a new era of federal involvement in controlling access to advanced AI systems before they become widely available.

What Are Sol, Terra, and Luna?

Altman described the three models as major advances in both capability and cost efficiency. Sol is positioned as a smart and efficient model that represents a clear upgrade in capability, priced at the same level as GPT-5.5 but with improved performance across several areas. Terra is part of the GPT-5.6 family and delivers performance similar to GPT-5.5 but at half the cost, making it significantly more affordable for developers and enterprises.

Luna rounds out the trio as the third model in the GPT-5.6 lineup. According to OpenAI, Sol shows particular improvements in coding and biology tasks and is the company's strongest model to date for cybersecurity-related work, helping developers identify and fix security vulnerabilities in software while maintaining OpenAI's internal safety standards.

Why Is the U.S. Government Blocking Wide Release?

The limited rollout stems from President Donald Trump's executive order earlier this month encouraging AI companies to submit their most advanced models for voluntary federal review before public launch. OpenAI has complied by sharing information about Sol, Terra, and Luna with U.S. authorities as part of early evaluation efforts.

According to an internal memo from Altman, federal authorities will approve access to the models "customer by customer" during an initial preview period, with a broader public release expected a few weeks later. Multiple U.S. agencies are involved in shaping this review process, including the Office of the National Cyber Director, the Office of Science and Technology Policy, and the Department of Commerce.

This approach mirrors recent restrictions placed on rival AI company Anthropic, which suspended access to two of its latest models after receiving a federal directive blocking availability to foreign nationals on security grounds.

How to Understand the Government's Role in AI Releases

• Voluntary Review Process: While Trump's executive order describes the reviews as voluntary, the practical enforcement suggests federal oversight is becoming mandatory in practice, with agencies actively controlling which customers receive early access to advanced models.
• Customer-by-Customer Approval: Rather than a blanket public release, the government will evaluate and approve individual customers or organizations before they gain access to Sol, Terra, and Luna during the preview phase.
• Temporary Measure: OpenAI has stated it does not view this arrangement as a long-term solution and is working with officials and industry partners to establish a more sustainable process for future AI releases.

What Does Altman Think About Government Control?

Altman has expressed a nuanced position on the government's involvement. According to reports, he acknowledged that extensive AI safety testing is valuable but opposed the government choosing which customers receive early access. OpenAI stated in its official announcement that it "supports wider access to its systems" and is working to make Sol, Terra, and Luna available to more users in the coming weeks.

"We don't believe this kind of government access process should become the long-term default," OpenAI said in its statement about the limited rollout.

OpenAI, Official Statement

The company emphasized that current access restrictions can delay availability for developers, enterprises, and security teams who depend on such tools. OpenAI is simultaneously working with the Trump administration to build a more structured system for reviewing advanced AI models before release, aiming to create a repeatable process that balances safety checks with faster availability.

The Broader Context: A Decade-Long Rivalry Shaping AI's Future

The government's new oversight approach comes amid a complex landscape of competition and conflict within the AI industry. According to reporting by the Wall Street Journal, there has been a decade-long feud between Altman and Dario Amodei, CEO of Anthropic, that has shaped how these companies approach AI development and safety.

Amodei left OpenAI in 2020 due to differences in business direction and founded Anthropic as what he positioned as a "healthy alternative to AI competitors." The conflict between the two leaders dates back to 2016 and centers on fundamental disagreements about how AI should be developed and governed. Amodei has emphasized safe and responsible AI development, while Altman has pursued a more aggressive commercialization strategy.

This rivalry became particularly visible recently when Anthropic refused to remove safety measures during contract negotiations with the U.S. Department of Defense, while OpenAI accepted a defense contract while maintaining similar safety standards. Amodei criticized OpenAI's public claims about its safety measures as "a blatant lie," highlighting the deep philosophical divide between the two companies.

When Will These Models Be Available?

OpenAI has not provided a fixed timeline for when Sol, Terra, and Luna will reach general availability, only stating that it expects them to become widely accessible "in the coming weeks". The company has also not confirmed which partners currently have access to the restricted preview. Altman responded to user questions about global availability by saying OpenAI is "working hard for worldwide" access, but no specific date has been announced.

Altman

The government's approval process adds uncertainty to the rollout timeline. Each customer or organization seeking access will need individual federal approval, potentially creating delays for some users while others gain earlier access based on their approval status.

What This Means for the AI Industry

The government-controlled preview represents one of the first major AI launches in which access is managed through federal oversight before wider availability. This development raises questions about whether the U.S. government's AI review process will become standard practice for all advanced model releases, fundamentally changing how AI companies bring products to market.

The approach suggests a shift toward more active federal involvement in AI governance, moving beyond voluntary guidelines toward practical control mechanisms. As AI systems become more powerful and capable, regulators appear increasingly willing to intervene in product release decisions, setting a precedent that could influence how other AI companies manage future launches.