gpt-5.6 pricing and cost-to-serve. cost-to-serve is what one user costs you every time they touch your ai feature. gpt-5.6's three tiers let you split those calls, so the cheap ones stop paying flagship prices. this is a product owner's read on the bill, not an engineer's read on the model.
on june 26, 2026, OpenAI shipped the gpt-5.6 family: sol, terra, and luna. luna runs at $1 in / $6 out per 1m tokens. sol runs at $5 / $30. that is a 5x spread on the same product surface, reported by finout.
if you lead product at a series b saas with a customer-facing ai feature, that spread is your next finance review. the question is not which model is smartest. the question is which calls actually need sol, and which have been quietly overpaying for it.
#what is gpt-5.6's luna tier actually for?
what is the gpt-5.6 luna tier and how cheap is it?
luna is the cheapest of openai's three gpt-5.6 tiers, launched june 26, 2026 at $1 input and $6 output per 1m tokens versus sol at $5 / $30 — a 5x gap. openai aims luna at classification, intent routing, and summarization: the high-volume, low-reasoning calls behind most ai features.
most customer-facing ai features are not one job. they are a pile of small jobs wearing one button. tag this message. route it to the right flow. summarize the thread. then, sometimes, actually reason about a hard case.
the first three are exactly what luna is priced for. the fourth is what you keep sol for. if every one of those calls currently hits your flagship model, roughly 80% of your token volume is paying 5x more than it needs to.
#how much does the tier split actually take off the bill?
take a feature doing 10m output tokens a month, where 80% of that is routing, tagging, and summarizing. on sol alone, that output costs $300k a month. move the 80% to luna and the same work costs $88k. the reasoning calls stay on sol, untouched, because that is where the quality has to hold.
the user sees no difference. the tag still lands, the summary still reads, the hard question still gets the good answer. the only thing that changed is which tier billed for which call.
#does caching change the math too?
yes, and it stacks. gpt-5.6 keeps a 90% discount on cached input, with explicit cache breakpoints and a 30-minute minimum cache life, per eesel on june 26, 2026. cache writes bill at 1.25x, so the one-time cost is small.
your system prompt is the same on every call. that is the part you cache. for a feature sending a 2k-token instruction block on every request, caching it takes 90% off that slice of input, on top of whatever tier you routed the call to.
we are
stennir is an ai consultancy that ships the routing layer so your feature reads each call and sends cheap work to luna and real reasoning to sol.
we aren't
stennir is not a reseller that swaps your model id to the cheapest tier and hopes the quality survives.
#do you have to rebuild the feature to get this?
no. our ai for product builds already split traffic by task. routing, tagging, and moderation go to the cheap tier. reasoning goes to the flagship. so a new tier like luna is a config change on the cheap lane, not a rebuild of the feature your users depend on.
if your feature sends everything to one model today, that is the gap to close first. the split is a day or two of work. the saving compounds every month your usage grows.
our cfo did not ask which model we used. he asked what one active user costs us. the tier split is the answer that let the feature survive the review.
this sits in our consultancy practice, and the journal has the receipts from features we have already taken to production. if you want your own numbers before the next finance review, book a 30-min discovery call and we will walk your feature's traffic split with you.