ChatGPT Users Are Getting a Major Upgrade This Week With Three New AI Models

• Major upgrade for ChatGPT users as OpenAI is set to release its new GPT-5.6 AI model after a weeks-long delay prompted by a U.S. government security review.

• The company says the update is its most advanced AI system yet, with major improvements in reasoning, cybersecurity and autonomous task execution.

OpenAI is preparing to launch GPT-5.6, its latest artificial intelligence model, following a delay caused by a U.S. government review over national security concerns.

According to Axios, the model was originally scheduled for release in June but was temporarily held back after the White House requested additional technical testing to assess potential cybersecurity and national security risks.

After weeks of discussions with OpenAI executives, U.S. officials have now cleared the model for public release.

OpenAI describes GPT-5.6 as its most powerful AI model yet, featuring significant improvements in reasoning, cybersecurity, biology and agentic AI—systems capable of carrying out complex tasks with minimal human supervision.

Why the Launch Was Delayed

The delay was not due to technical issues but growing concerns that increasingly powerful AI systems could be exploited by cybercriminals or foreign governments.

Officials were particularly concerned that advanced AI models could identify software vulnerabilities that hackers or intelligence agencies might use for cyberattacks.

A similar review recently affected AI company Anthropic, whose advanced models also faced temporary restrictions before eventually receiving approval.

Although OpenAI cooperated with the review, the company said it does not want government approval to become a permanent requirement for future AI releases.

“We don’t believe this kind of government access process should become the long-term default because it keeps the best tools away from users, businesses and researchers,” the company said.

Three Versions of GPT-5.6

OpenAI says the GPT-5.6 family will launch in three versions:

• Sol – The flagship model with the highest reasoning and performance capabilities.
• Terra – A balanced version designed for everyday use.
• Luna – The fastest and most affordable model, optimized for speed and high-volume tasks.

The company also said GPT-5.6 incorporates a layered safeguard system designed to reduce the risk of misuse for cyberattacks and other harmful activities.

New U.S. AI Security Framework

The release comes amid a new U.S. government approach to regulating advanced AI systems.

In June, President Donald Trump signed an executive order establishing a voluntary framework that allows AI companies to submit new models to the government for review for up to 30 days before public release.

OpenAI said it is working with U.S. officials to develop a repeatable security review process for future model launches.

For millions of ChatGPT users—including students, developers, businesses and professionals—the launch of GPT-5.6 is expected to bring faster performance, stronger reasoning capabilities and more reliable responses across a wide range of tasks.