Less than 24 hours after news broke that OpenAI would delay its next model release at the Trump administration’s request, the company went and released it anyway. GPT-5.6 is here.
The new model suite comes in three flavors: Sol, the flagship powerhouse; Terra, a mid-tier option for high-volume workloads; and Luna, the budget-friendly everyday model. OpenAI says the whole suite is particularly good at coding, cybersecurity, biology, and long-horizon agentic tasks.
Pricing for Sol sits at $5 per million input tokens and $30 for output — roughly half the cost of Anthropic’s Claude Fable 5. Terra and Luna are cheaper still, making the full suite competitive on price.
The timing is interesting. The Trump administration had reportedly asked OpenAI to stagger the release, citing safety concerns. OpenAI appears to have said “thanks but no thanks” and pushed forward with the limited preview anyway.
Whether this is a principled stand for open AI development or just good old-fashioned corporate defiance, it sets up a fascinating tension between government regulators and the companies building the future of AI.
