OPENAI
GPT-5.6 passes the sniff test
OpenAI is set to launch GPT-5.6, after delaying the release due to US government security concerns.
Officials were worried that powerful AI models could be misused for advanced cyberattacks, especially against older and more complex systems.
The US and China are both increasing scrutiny around frontier AI, as access to these models becomes a bigger national security issue.
According to Axios, the Trump administration approved a wider GPT-5.6 launch after extra testing and meetings with OpenAI.
The company had previously limited access to a small group of vetted partners.
The bigger picture:
Governments are taking a closer look at major AI launches.
Cybersecurity is becoming one of the biggest concerns.
AI companies are being pushed to prove their safeguards actually work.
Frontier models watched
OpenAI will release its top model, GPT-5.6 Sol, alongside two cheaper versions, Terra and Luna.
The company says the models have stronger abilities in coding, biology and cybersecurity.
Other AI companies are facing similar pressure.
Anthropic recently restricted access to its advanced models after a US export control order, though some limits have since been lifted after new safeguards were added.
GPT-5.6 didn’t get a launch delay. It got summoned to the principal’s office by national security.- MG