Home / GPT-6 Luna Cost Calculator
GPT-6 Luna is the lowest-priced GPT-6 tier — the one you pick when request volume is huge and per-request cost has to round to fractions of a cent. Type in your token counts below to see what Luna costs per request, per day, per month, and per year, with prompt caching and service tiers included. The widget also shows Sol and Astra on the same inputs, so you can watch the price gap scale with your volume.
Pricing source: OpenAI official pricing · Last verified: October 10, 2026 · Manually verified
OpenAI's official pricing page lists two rate columns for GPT-6 Luna — short context and long context. Both are shown here; the calculator selects the correct column above 272,000 input tokens.
| Rate (per 1M tokens) | Short context | Long context |
|---|---|---|
| Input | $0.10 | $0.20 |
| Cached input | $0.01 | $0.02 |
| Output | $0.50 | $0.75 |
Two things shape every Luna bill. First, cached input costs one-tenth as much as fresh input ($0.01 vs $0.10) — the same 0.1× cache ratio as its siblings, at one-twentieth of Sol's absolute rates — so workloads with repeatable prompt prefixes — system prompts, few-shot examples, retrieved context — see their input line shrink fast. Second, a note on shape: output costs five times the input rate here — the same 5× ratio as Sol and Astra, so Luna's bill is not shaped differently, it is simply smaller. At $0.50 per 1M output tokens (one-twentieth of Sol's $10.00), a generation-heavy workload costs 20× fewer dollars at identical token counts. Service-tier multipliers are the same across all GPT-6 models: Standard 1.0×, Flex 0.5×, Batch 0.5× — an official 50% discount for deprioritized or deferred processing.
1,500 input + 300 output tokens per request, 5,000,000 requests/month, 80% prompt caching.
$0.000192 / request · $960 / month
fresh input 300 × $0.10 + cached 1.2K × $0.01 + output 300 × $0.50 (per 1M)
4,000 input + 800 output tokens per request, 500,000 requests/month, 70% prompt caching.
$0.000548 / request · $274 / month
fresh input 1.2K × $0.10 + cached 2.8K × $0.01 + output 0.8K × $0.50 (per 1M)
Identical workload (4K in / 0.8K out, 500K requests/month, 70% cache), switched from Standard to Batch.
$137 / month — the official 0.5× multiplier, applied
Batch = 24-hour-turnaround deferred processing. Not for latency-sensitive traffic.
All examples use short-context rates, last verified October 10, 2026. Tier multipliers are OpenAI's documented discounts.
At short-context rates, GPT-6 Luna costs $0.10 per 1M input tokens, $0.01 per 1M cached input tokens, and $0.50 per 1M output tokens. Long-context rates are $0.20 input, $0.02 cached input, and $0.75 output. The calculator automatically applies long-context rates above 272,000 input tokens for the full request.
Very cheap at scale. A high-volume classification workload — 1,500 input and 300 output tokens per request, 5 million requests a month, 80% prompt caching — costs $0.000192 per request, under two-hundredths of a cent. At that volume the monthly bill is $960.
Cached input on Luna costs $0.01 per 1M tokens versus $0.10 for fresh input — one-tenth the price on the cached share. A support chatbot doing 500,000 requests a month with 70% cache hits pays $74/month for input tokens instead of $200, bringing the total bill from $400 down to $274.
Yes, on paper: Batch is an officially documented 0.5× multiplier — a 50% discount for deferred, 24-hour-turnaround processing. The $274/month support-chatbot example would drop to $137/month. The tradeoff is latency and scheduling: Batch is not for traffic that needs answers now.
OpenAI's official pricing page lists two rate columns for Luna: short context ($0.10 input / $0.01 cached / $0.50 output per 1M tokens) and long context ($0.20 input / $0.02 cached / $0.75 output). Input exactly doubles at long context; output rises by half. The calculator applies long-context rates automatically above 272,000 input tokens.
Purely on price, Luna's rates are one-twentieth of Sol's: $0.10 vs $2.00 per 1M input tokens and $0.50 vs $10.00 per 1M output tokens. A 500,000-request chatbot workload costs $274/month on Luna versus $5,480/month on Sol. See the full side-by-side breakdown on the Sol vs Luna cost comparison page. This is a mathematical cost comparison only — it says nothing about which model to choose.
All GPT-6 Luna prices on this page are copied from the OpenAI official pricing page. Last verified: October 10, 2026 · Manually verified — there is no live price feed, and we don't pretend there is. The full policy, including which rate tables the calculator uses and why, is published on the Pricing Methodology page.
Independence disclosure: this is an independent cost-estimation tool. It is not affiliated with, sponsored by, or endorsed by OpenAI. Model names and prices belong to their providers — always confirm on the official pricing page before making decisions.