OpenAI GPT 5.6 Price Cut

OpenAI just cut the price of its cheapest GPT-5.6 model by 80 percent. Three weeks after launching it.

The Luna tier dropped from $1/$6 to $0.20/$1.20 per million input/output tokens. Terra took a 20 percent cut too. The flagship Sol model held steady at $5/$30, for now.

The trigger isn't hard to find. A CNBC investigation this week showed Chinese models have captured 46 percent of US enterprise token usage on OpenRouter, with DeepSeek V4 Pro running promotional pricing as low as $0.435/$0.87 per million tokens. Frontier labs built their economics on the assumption that quality would command a premium. That assumption is now being tested in real time, and the answer so far is: not for long.

For enterprise buyers, this is the signal to stop treating model choice as a fixed line item. Pricing is moving faster than most procurement cycles, and the labs that win won't be the ones with the best benchmark score, they'll be the ones you can swap out without rebuilding your stack.

If frontier model prices keep compressing this fast, where does the real value settle, the model itself or what you build on top of it?