OpenAI dropped prices on two of its own models this week, and the cuts aren’t small. GPT-5.6 Luna, the budget tier, is down 80 percent, from a dollar to twenty cents per million input tokens. GPT-5.6 Terra is down 20 percent. Sol, the top tier, keeps its price but gets a new Fast mode that runs up to two and a half times quicker if you’re willing to pay double for the speed.
Luna launched at that higher price only a few weeks ago. Cutting it 80 percent this fast tells you the price war between AI labs isn’t slowing down, it’s speeding up. Right around the same time, DeepSeek pushed its own flash model into public beta at a price low enough to undercut OpenAI’s new discount anyway.
Here’s what I like about this moment. A year or two back, the AI pricing conversation was mostly about scarcity, who has enough compute, who can even afford to serve you. Now it’s shifting toward abundance. Multiple labs are fighting to be the cheapest capable option, and the tool getting cheaper is the same tool getting more capable at the same time.
That’s the part that hits hardest for the builder who isn’t backed by a big company. A solo developer or a small startup can now run a serious AI feature in production without the token bill wrecking their month. That changes who gets to build things, not just what gets built.
The skeptical read is that these price wars are unsustainable and someone eventually has to make the math work. Probably true. But while it lasts, the person who benefits most is the one actually using these tools to build something, not the labs racing each other to the bottom.
——
Follow: @Ali Demi
Book your free AI clarity call, NOW!
https://buff.ly/TpWy277
Sources:
https://www.cnbc.com/2026/07/30/open-ai-price-cut-gpt.html
https://venturebeat.com/technology/ai-price-wars-openai-cuts-gpt-5-6-luna-prices-by-80-as-model-competition-shifts-toward-cost
https://www.axios.com/2026/07/30/openai-cuts-prices-gpt-terra-luna5
Repost this. Thanks.

