OpenAI Cuts GPT-5.6 Luna Price by 80% and Terra by 20%

My take: OpenAI just cut prices on two models from the GPT-5.6 family, only three weeks after launch. GPT-5.6 Luna, the fastest and most affordable model, drops 80%: it now costs $0.20 per million input tokens and $1.20 per million output tokens. GPT-5.6 Terra, the balanced model for everyday work, drops 20%: $2 and $12 per million tokens, respectively.

This matters because competition between AI labs is pushing prices down at a pace that two years ago seemed unthinkable. Every time inference costs drop like this, something that was expensive or out of reach for a small business or freelancer becomes accessible. I track these prices myself because they directly affect what I can build and at what cost.

If you still don't have a strategy for integrating AI models into your workflow, the cost argument no longer holds. The question now is: what process or product could you improve if the cost of AI stopped being a barrier?