The tech world never stops, but it looks like Washington just pulled the handbrake on OpenAI. Just hours after rumors suggested that the government wanted to stall things, OpenAI went ahead and launched its next big thing: the GPT-5.6 family of AI models. This release comes with a twist, though. Instead of the usual numbers, we are getting a cool space-themed family of three GPT-5.6 models: Sol (the powerhouse), Terra (the everyday worker), and Luna (the fast, budget-friendly one).
OpenAI’s official announcement shows that these models are seriously good at coding, biology, and cybersecurity. The flagship model, Sol, even gets a “max reasoning” feature and an “ultra” mode that lets multiple sub-agents team up to crush complex tasks. It is basically OpenAI flexing its muscles.
An absolute price war
If the tech upgrades don’t turn heads, the pricing definitely will. OpenAI is throwing down the gauntlet and starting a massive price war. They are offering the flagship Sol model at $5 per million input tokens and $30 per million output tokens.
To put that into perspective, that is almost half the cost of Claude Fable 5. Anthropic’s latest AI model charges a hefty $10 input and $50 output. The mid-tier Terra drops that price in half, and Luna is even cheaper. OpenAI is making it crystal clear that they want to make powerful AI accessible without breaking the bank.
The White House locks the gates
Now for the bad news: you won’t be jumping into ChatGPT to play with Sol today. In a notable twist, the Trump administration has put this launch on a tight leash. For now, GPT-5.6 is stuck in a limited preview. Only about 20 companies—all hand-picked and approved case-by-case by the government—get to try it out. This shows just how serious Washington is getting about advanced AI access.
This gatekeeping comes right after federal regulations forced Anthropic to completely pull its public access to Claude Fable 5 and Mythos 5 over safety fears. Washington fears that these new models are too good at hacking. Because of this, OpenAI spent a massive 700,000 GPU hours testing safety, ensuring Sol is geared toward fixing digital security flaws rather than causing chaos.
Play by the rules, for now
OpenAI CEO Sam Altman and his team have been playing nice with the White House for the past month. However, they didn’t hide their frustration in their latest blog post. The company openly admitted that while they are cooperating to keep things safe, they really don’t want this kind of government babysitting to become the permanent norm.
OpenAI argued that hiding the best tools behind red tape just hurts developers, businesses, and cyber defenders who actually need them. For now, the firm is treating this state-monitored rollout as a temporary compromise while they work with the administration to figure out a smoother, repeatable process for future updates.
The post OpenAI Drops GPT-5.6 Sol, Terra, and Luna with an Availability Catch under Tight Government Rules appeared first on Android Headlines.
