OpenAI Cuts GPT-5.6 Luna Pricing by 80%, Reshaping the Cost-Per-Task Frontier
OpenAI slashed per-task pricing on its GPT-5.6 Luna model by 80% and Terra by 20%, a move that reads less like a promotion and more like a shot in an intensifying cost-efficiency war.
OpenAI cut pricing across its GPT-5.6 lineup on Thursday, and the numbers are steep enough to change how teams budget for production workloads. According to @ArtificialAnlys, the Luna variant is now roughly 80% cheaper per task, while Terra received a more modest 20% reduction in cost per task. The distinction matters: a headline token-price cut and a real cost-per-task cut are not the same thing, and Artificial Analysis framed the change specifically in the latter terms.
The framing around the Pareto frontier is the part builders should read carefully. Artificial Analysis noted that Terra "remains largely behind Luna and Sol" on the cost-versus-capability curve even after its price drop. In plain terms, Terra got cheaper but did not become the obviously correct choice for most tasks — Luna and Sol still dominate the frontier where you get the most capability for the least money. A 20% cut to a model that was already off the frontier doesn't move it onto the frontier.
Get our free daily newsletter
Get this article free — plus the lead story every day — delivered to your inbox.
Want every article and the full archive? Upgrade anytime.
No spam. Unsubscribe anytime.