OpenAI on Thursday officially launched GPT-5.6, its most advanced artificial intelligence model, following a delay prompted by U.S. government scrutiny over national security concerns surrounding the release of increasingly powerful AI systems.
The launch comes amid intensifying global competition between the United States and China to develop next-generation AI models capable of advancing fields such as cybersecurity, scientific research, and software development.
U.S. authorities have stepped up oversight of frontier AI models amid concerns that advanced systems could be exploited by foreign military and intelligence agencies or used to accelerate sophisticated cyberattacks targeting critical infrastructure.
The release follows additional testing and consultations between OpenAI and U.S. government officials before approval was granted for a broader public rollout.
OpenAI had previously limited access to GPT-5.6 to a select group of vetted partners while the review process was underway.
Alongside its flagship GPT-5.6 Sol model, the company also introduced two lower-cost variants, Terra and Luna, expanding its portfolio of AI models for different use cases.
According to OpenAI, GPT-5.6 delivers improved performance in coding, biology, cybersecurity, reasoning and autonomous AI tasks. During earlier previews, the company said GPT-5.6 Sol demonstrated competitive performance on cybersecurity benchmarks against leading frontier AI models.
The launch comes as governments increasingly examine the security implications of advanced AI technologies. Chinese authorities have also been considering restrictions on overseas access to the country’s most advanced AI models as Beijing seeks to safeguard strategic AI capabilities.
The rollout marks OpenAI’s biggest model release since GPT-5 and reinforces the company’s position in an increasingly competitive AI market, where rivals including Anthropic, Google, Meta and Elon Musk’s xAI continue to introduce more capable frontier models.
