OpenAI Launches GPT-5.6 Sol
OpenAI just dropped GPT-5.6 Sol at half the price of its main rival! 🤯 But there's a huge catch with the government... Tap to watch!
What the video says
Nearly half the cost of its main rival, OpenAI's brand-new GPT-5.6 model suite has officially arrived on Friday. This sudden release comes less than 24 hours after news broke that the company would stagger its rollout at the request of the Trump administration.
The new flagship model, SOL, is priced at $5 input and $30 output per million tokens, which commentators note is a sharp undercut to Anthropic's CLAUDE Fable-5 at $10 input and $50 output. Alongside the flagship, OpenAI unveiled Terra for high-volume work at half of Sol's cost and Luna as a fast everyday model.
The launch occurs under unprecedented government oversight, with the Trump administration set to approve preview customers on a case-by-case basis. To address security concerns in Washington, D.C., OpenAI dedicated approximately 700,000 A100-E GPU hours to automated red teaming.
The company explicitly stated that GPT-5.6 is trained to refuse prohibited cyber assistance, including when users attempt to disguise their intent or jailbreak the model. Furthermore, the company claims SOL does not cross the cyber-critical threshold under its preparedness framework, which was recently revised in April to remove some previous study areas.
While OpenAI is cooperating with the White House for this restricted preview, they have made their resistance to permanent oversight clear. By the company's own count, this level of state intervention is a temporary compromise rather than a permanent new standard.
The firm stated they do not believe this kind of government access process should become the long-term default, because it keeps the best tools from global partners. The restricted preview will run for the next 2 weeks with third-party testers, ahead of a planned general release in the coming weeks.
