OpenAI · 2026-07-30 · major
GPT-5.6 Luna cut 80%, Terra 20% — OpenAI drops API prices three weeks after launch
OpenAI cut GPT-5.6 Luna to $0.20 / $1.20 per million input and output tokens (80% off) and GPT-5.6 Terra to $2 / $12 (20% off) on July 30, 2026. Sol pricing is unchanged. Serving costs fell 20% and token efficiency rose 15%.

OpenAI drops GPT-5.6 Luna 80% and Terra 20% three weeks after launch, passing on 20% serving-cost gains.
Key specs
| Luna reduction | 80% |
|---|---|
| Terra reduction | 20% |
| Serving cost drop | 20% |
| Token generation efficiency gain | 15% |
Quick facts
| Maker | OpenAI |
|---|---|
| Models repriced | GPT-5.6 Luna, GPT-5.6 Terra |
| Luna price | $0.20 / 1M in · $1.20 / 1M out (was $1 / $6) |
| Terra price | $2 / 1M in · $12 / 1M out (was $2.50 / $15) |
| Sol price | Unchanged |
| Effective date | July 30, 2026 |
| How they cut it | 20% lower serving cost, 15% better token efficiency |
Pricing
| GPT-5.6 Luna — input · was $1 | $0.20 / 1M tokens |
|---|---|
| GPT-5.6 Luna — output · was $6 | $1.20 / 1M tokens |
| GPT-5.6 Terra — input · was $2.50 | $2.00 / 1M tokens |
| GPT-5.6 Terra — output · was $15 | $12.00 / 1M tokens |
What is it?
GPT-5.6 Luna is now $0.20 per million input tokens and $1.20 per million output tokens on the OpenAI API — an 80% cut. GPT-5.6 Terra, the balanced tier, drops to $2 in / $12 out — 20% cheaper. Sol, the flagship, keeps its old price. All three ship as part of the GPT-5.6 family that OpenAI launched July 9, 2026.
How does it work?
OpenAI attributes the cut to infrastructure work done since launch: the company says end-to-end serving costs for the GPT-5.6 family fell 20% and token-generation efficiency climbed more than 15%. Those gains land on the two cheaper tiers first — Luna at 80% off, Terra at 20% off — with Sol pricing held so the flagship keeps funding continued frontier training.
Why does it matter?
At $0.20 / $1.20, GPT-5.6 Luna is now cheaper than most open-weight-hosted models on OpenRouter and clearly undercuts Anthropic's Haiku 4.5 and Google's Gemini 3.6 Flash on paper. That reshapes the cheap-tier pricing race: high-throughput classifiers, routers, and background workflows that were the natural home of open-source models suddenly have a hosted OpenAI option at a similar unit cost.
Who is it for?
API builders running high-volume, cost-sensitive workloads on Luna or Terra
Frequently asked questions
- How much does GPT-5.6 Luna cost now?
- GPT-5.6 Luna now costs $0.20 per million input tokens and $1.20 per million output tokens on the OpenAI API, down from $1 and $6 — an 80% price cut. It is OpenAI's fastest and cheapest model in the GPT-5.6 family and is meant for high-throughput classification, routing, and small edits.
- How much does GPT-5.6 Terra cost now?
- GPT-5.6 Terra is now $2 per million input tokens and $12 per million output tokens, down from $2.50 and $15 — a 20% cut. Terra is OpenAI's balanced GPT-5.6 tier, priced to sit between the cheap Luna and the flagship Sol tier while keeping GPT-5.5-class quality.
- Did GPT-5.6 Sol get cheaper too?
- No. OpenAI kept GPT-5.6 Sol pricing unchanged on July 30, 2026. Only Luna and Terra were repriced. Sol Ultra with max reasoning stays at the top of OpenAI's coding-agent leaderboard and still costs less than half of Claude Fable 5 on the Artificial Analysis Coding Agent Index, per OpenAI.
- Why is OpenAI cutting GPT-5.6 prices so soon after launch?
- OpenAI said serving-cost improvements dropped the end-to-end cost of running the GPT-5.6 family by 20% and lifted token-generation efficiency by more than 15% in the three weeks since launch. The company is passing those gains on to Luna and Terra customers to expand adoption on cheaper tiers.
- When does the new GPT-5.6 pricing take effect?
- The new GPT-5.6 Luna and Terra prices are live on the OpenAI API pricing page as of July 30, 2026 — the same day OpenAI announced the cut. No opt-in or contract change is required; every API call from that day forward is metered at the lower rate.
Try it
gpt-5-6-luna and gpt-5-6-terra on the OpenAI API — new prices billed automatically