OPENAI JUST ANNOUNCED A 50% API PRICE CUT. REOPEN THE PROJECTS YOU COULDN'T AFFORD TO RUN. On September 22, OpenAI introduced GPT-6 Sol and GPT-6 Luna, bringing advances from Astra into models it describes as faster and more affordable. The pricing comparison matters: 50% lower than GPT-5.6 promotional pricing. That is the stated baseline, not a promise that every AI workflow now costs half as much. OpenAI says: "we're passing the savings directly to you". For builders, the interesting question is what becomes worth running at the new rates. A document-processing job. A support triage step. A research pipeline you shelved because the volume made it too expensive. Take one of those projects off the shelf. Replay a representative batch on the new model. Keep the inputs and acceptance criteria fixed. Measure how many outputs pass review and what the entire batch costs, including retries and human corrections. If quality holds and the total cost falls, you have evidence to reconsider the project. If review work increases, include that before calling it a saving. A price cut can move the line between a useful demo and a service you can afford to operate. It does not tell you where that line sits for your product. Recalculate the idea you rejected. The old spreadsheet may be ruling out a viable business.