Pricing · OpenAI

GPT-5.6 Luna

OpenAI's gpt-5.6-luna, 1.1M tokens of context. Every number on this page was read off OpenAI's own pricing page — never estimated, never recalled from memory.

How OpenAI describes it

“GPT-5.6 Luna is designed for cost-sensitive, high-volume workloads.”

Pricing

Prices in US dollars per 1M tokens, as OpenAI's model page printed them when they were last verified, on November 16, 2026.

Input $0.20 per 1M tokens
Output $1.20 per 1M tokens
Cached input $0.02 per 1M tokens · 0.1× the input price
Cache write (30-minute retention) $0.25 per 1M tokens
Cache write (12-hour retention) $0.50 per 1M tokens

Which cache write is which: OpenAI's pricing page in its Markdown version, read October 29, 2026, heads its Standard table "| Model | Short context input | Short context output | Short context cached input | Short context 30m write | Short context 12h write | Long context input | Long context output | Long context cached input | Long context 30m write | Long context 12h write |" and prints "| gpt-5.6-luna | $0.20 | $1.20 | $0.02 | $0.25 | $0.50 | $0.40 | $1.80 | $0.04 | $0.50 | $1.00 |".

Long-prompt tiers

Pricing changes above a prompt-size threshold. If your average request is near the boundary, the effective rate is not the headline rate.

Condition Input Output Cached input Cache write (30-minute retention) Cache write (12-hour retention)
prompt > 272000 tokens $0.40 $1.80 $0.04 $0.50 $1.00

At the calculator's starting volume, 1M input tokens and 200K output tokens with no cache hits, GPT-5.6 Luna costs $0.44: $0.20 for input and $0.24 for output, at the prices above. Same formula as the calculator; change the volume there.

Where this price sits

$0.20 per 1M input tokens (for prompts up to 272K tokens) is the 14th cheapest input price of the 82 current models with a verified price on AI Signal, and the 2nd of OpenAI's 20. Output costs 6.0× the input price.

Other current models in the GPT-5.6 family:

Other price lists OpenAI publishes for GPT-5.6 Luna

The prices above are OpenAI's Standard list, the one the table and the calculator use. OpenAI also sells GPT-5.6 Luna on 3 other lists, each with a condition. Figures and wording are quoted from OpenAI's own pages as read on the date shown; AI Signal computes none of them.

Mentioned in

3 posts on AI Signal cite GPT-5.6 Luna. The sentence is the post's own, taken from where it links here.

Recorded price changes

Every change AI Signal has recorded for GPT-5.6 Luna, with the date it took or takes effect and why. Corrections of our own reading are listed too, and say so.

Every price change recorded on AI Signal →

Specifications

GPT-5.6 Luna's API identifier, provider and limits, from the same row of the dataset as the prices above.

API model id gpt-5.6-luna
Provider OpenAI
Family GPT-5.6
Context window 1.1M tokens
Max output 128K tokens
Status Current

It returns up to 128K tokens per response, against a context window of 1.1M tokens.

Questions about GPT-5.6 Luna pricing

How much does GPT-5.6 Luna cost per million tokens?
$0.20 per 1M input tokens and $1.20 per 1M output tokens for prompts up to 272,000 tokens, read on OpenAI's model page on November 16, 2026.
What does a cached input token cost on GPT-5.6 Luna?
$0.02 per 1M cached input tokens, 10% of the $0.20 input price; a 30-minute cache write costs $0.25, a 12-hour cache write costs $0.50 per 1M tokens.
Does GPT-5.6 Luna cost more for long prompts?
Yes. For prompts over 272,000 tokens it costs $0.40 per 1M input tokens and $1.80 per 1M output tokens, $0.04 for cached input, against $0.20 and $1.20 below that.

Verified by hand on 2026-09-24 against the model page this row cites, developers.openai.com/api/docs/models/gpt-5.6-luna, which prints 'Input $0.20 Cached input $0.02 Output $1.20' per 1M tokens, '1,050,000 context window', '128,000 max output tokens' and 'Prompts with >272K input tokens are priced at 2x input and 1.5x output for the full request. Cache writes are billed at 1.25x the uncached input token rate.' Until September 22, 2026 this row cited OpenAI's pricing page. That day, with the release of GPT-6 Sol and GPT-6 Luna, OpenAI reorganised the page: a 'Flagship models' table with the GPT-6 family and a 'Cyber models' table with GPT-5.6 Sol and Cyber, and this model no longer appears on it. It is not deprecated: OpenAI's deprecations page does not list it, its model page still prices it and OpenRouter still serves it; the citation moved to the model page so the weekly check keeps verifying the figure. Correction, October 26, 2026: the sentence above saying that this model no longer appears on OpenAI's pricing page rested on a reading of OpenAI's pricing page that did not see its folded 'All models' table, which the page draws in the browser and its Markdown version (developers.openai.com/api/docs/pricing.md) prints. Read there on 2026-10-26, its Standard table prints 'gpt-5.6-luna | $0.20 | $1.20 | $0.02 | $0.25 | $0.50 | $0.40 | $1.80 | $0.04 | $0.50 | $1.00' (short context input, output, cached input, 30m cache write, 12h cache write, then the same for long context, per 1M tokens). These are the same figures as this row; the row keeps citing its model page, and its Batch, Flex and Fast mode figures, removed from this site's lists on 2026-09-24, are back (secondPriceLists.json).

Verified November 16, 2026 against OpenAI's model page. Confidence: confirmed. How the data is checked.

Other OpenAI models with a verified price

Cheaper or equally priced alternatives

Current models with an input price at or below $0.20 per 1M tokens.