Home › Models › GPT-5.6 Luna
GPT-5.6 Luna API pricing
GPT-5.6 Luna is a model from OpenAI with a 1,050,000-token context window. Rates below are per 1,000,000 tokens, verified October 2026.
Price list
| Token class | Rate / 1M | Notes |
|---|---|---|
| Input | $0.20 | Prompt tokens not served from cache |
| Cached input | $0.02 | 90% below base input |
| Cache write | $0.25 | Charged once when the prefix is stored |
| Output | $1.20 | 6.0× the input rate |
| Long context input | $0.40 | Above 272,000 input tokens |
| Long context output | $1.80 | Same threshold |
| Batch API | −50% | Input and output, results within 24 h |
Budget tier of the 5.6 family, sold alongside GPT-6 Luna.
What GPT-5.6 Luna costs at three real workloads
Sticker prices rarely match the invoice. These three workloads use the rates above, with cache writes charged once per job and the batch discount applied only where the provider offers it.
| Workload | Requests | Tokens | Estimated cost |
|---|---|---|---|
| Support chat | 1,000 | 4K in / 600 out | $1.52 |
| RAG with caching | 1,000 | 50K in (45K cached) / 1K out | $3.11 |
| Nightly batch job | 10,000 | 10K in / 2K out | $22 |
Caching the 45K-token prefix saves about 72% on the RAG workload versus sending it uncached every call ($11.20 → $3.11). Batch mode halves the nightly job.
Where GPT-5.6 Luna sits in the market
Out of 28 models in this dataset, GPT-5.6 Luna ranks #2 cheapest by input rate, #2 by output rate, and #2 on a blended 1M-input-plus-250K-output workload.
| Model | Provider | Input | Output | Blended |
|---|---|---|---|---|
| GPT-5.6 Luna | OpenAI | $0.20 | $1.20 | $0.50 |
| GPT-6 Luna | OpenAI | $0.10 | $0.50 | $0.225 |
| GPT-5.4 nano | OpenAI | $0.20 | $1.25 | $0.512 |
| DeepSeek V4.1 Flash | DeepSeek | $0.30 | $1.20 | $0.60 |
Blended cost = 1M input tokens + 250K output tokens at list price.
FAQ
How much does GPT-5.6 Luna cost per 1M tokens?
$0.20 per 1M input tokens and $1.20 per 1M output tokens on the standard tier, rising to $0.40 / $1.80 above 272,000 input tokens.
Is prompt caching cheaper on GPT-5.6 Luna?
Yes. Cached input bills at $0.02 per 1M tokens, 90.0% below the $0.20 base input rate, with a one-off cache write at $0.25 per 1M.
Does GPT-5.6 Luna support the Batch API?
Yes. Batch requests bill at roughly half the standard input and output rates: the nightly job above costs $44 at list price and about $22 in batch mode.