OpenAI: GPT-5.6 Luna
gpt-5.6-luna- Input / Output-70%
- $0.20 / $1.20$0.06 / $0.36
- Context
- 1.05M
- Released
- Jun 26, 2026
- Max output
- 128K
- Modalities
- VisionChat
- Capabilities
- Tool usePrompt CacheReasoning
About GPT-5.6 Luna
GPT-5.6 Luna is the smallest and most cost-efficient model in OpenAI's GPT-5.6 family, built for fast, high-volume work where a response has to be quick and cheap rather than deeply reasoned. It reads text and images, calls tools, returns structured JSON, and offers adjustable reasoning. Choose Luna over its siblings Sol and Terra when volume matters more than peak capability.
Where it works well
- Sits at the low-cost, fast end of the GPT-5.6 family, which suits steady high-volume traffic.
- Accepts images next to text, so screenshots and scanned pages can go through the same request.
- Supports tool calling and structured JSON output, which keeps it usable as a step inside automated pipelines.
- Reasoning is adjustable, so simple turns can skip extended thinking while harder ones get more.
When to choose another model
- Sol and Terra handle long agentic runs and difficult coding better; Luna is the tier to leave when tasks keep failing.
- OpenAI now lists the GPT-5.6 family as legacy, pointing new projects to the GPT-6 models instead.
Getting started
Create API key
Create a key in Console, then use it with every model on the platform.
Send your first request
Copy the example for your language and run it against the endpoint.
Text to textResponses APIPOST/v1/responsesAPI format:curl https://api.tokenlab.sh/v1/responses \ -H "Content-Type: application/json" \ -H "Authorization: Bearer sk-xxx" \ -d '{ "model": "gpt-5.6-luna", "input": "Hello!" }'
Pricing
The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.
Input tokens <= 272K
per 1M tokens- Official price
- Input $0.20 / Output $1.20 / Cache read $0.02 / Cache write $0.25
- Verified price
- Input $0.06 / Output $0.36 / Cache read $0.006 / Cache write $0.075
- Discount
- -70%
Input tokens 272K-1.05M
per 1M tokens- Official price
- Input $0.40 / Output $1.80 / Cache read $0.04 / Cache write $0.50
- Verified price
- Input $0.12 / Output $0.54 / Cache read $0.012 / Cache write $0.15
- Discount
- -70%
Cache read
- Official price
- $0.02
- Verified price
- $0.006
- Discount
- -70%
Cache write
- Official price
- $0.25
- Verified price
- $0.075
- Discount
- -70%
Charged per call, only when the model uses the tool.
Web search
- Official price
- $0.01/search
- Verified price
- $0.003/search
- Discount
- -70%
| Official priceper 1M tokens | Verified priceper 1M tokens | Discount | |
|---|---|---|---|
| Input tokens <= 272K | Input $0.20 / Output $1.20 / Cache read $0.02 / Cache write $0.25 | Input $0.06 / Output $0.36 / Cache read $0.006 / Cache write $0.075 | -70% |
| Input tokens 272K-1.05M | Input $0.40 / Output $1.80 / Cache read $0.04 / Cache write $0.50 | Input $0.12 / Output $0.54 / Cache read $0.012 / Cache write $0.15 | -70% |
| Prompt cache pricing | |||
| Cache read | $0.02 | $0.006 | -70% |
| Cache write | $0.25 | $0.075 | -70% |
| Tool feesCharged per call, only when the model uses the tool. | |||
| Web search | $0.01/search | $0.003/search | -70% |
Usage & activity
Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.
Usage & availability
Last 24 hours- Requests
- Success rate
- P95 latency
- Total tokens
- 30-day success rate
- 93.2%
- 7-day median latency
- 6.2sn=149
- 7-day P95 latency
- 25.9sn=149
- 30-day requests
- 2K+
- Last active
- 2 hours ago
Data is based on aggregate user requests, excluding status checks.
Open in Console
Open GPT-5.6 Luna in Console with a prompt ready to edit or send.
Help me try gpt-5.6-luna with a short message at /v1/responses. Show the reply, latency, and cost.
Use cases
Best for- Code
- Reasoning
- Vision
Classification and tagging
Sort tickets, label content, or extract fields from large volumes of short inputs where each call must stay inexpensive.
Lightweight assistants
Power in-product helpers that answer short questions, rewrite text, or call a single tool without long deliberation.
Screenshot reading
Pull text and values from interface captures or receipts, then return them as schema-checked JSON.
Sub-agent steps
Let a larger model delegate narrow, repetitive subtasks to Luna to keep the total bill down.
Prompt examples
Classify each of these 200 support emails as billing, bug, or feature request and return a JSON array with the ticket number and label.
Read this screenshot of an invoice and return vendor, date, and total as JSON.
Rewrite this release note in two plain sentences for a customer newsletter.
This model has conditional pricing. Monthly token totals alone cannot produce a reliable estimate; use the detailed pricing for the request specification, cache, and applicable time window.
FAQ
What is GPT-5.6 Luna?
GPT-5.6 Luna is the small, cost-efficient tier of OpenAI's GPT-5.6 family, alongside Sol and Terra. It is meant for fast, high-volume workloads, with text and image input, tool calling, and structured output.
How does Luna differ from GPT-5.6 Terra and Sol?
Sol is the flagship for complex reasoning and long agentic work, Terra is the balanced middle, and Luna trades depth for speed and cost. Start with Luna for simple, repeated tasks and step up when quality drops.
Does GPT-5.6 Luna support images?
Yes. It accepts images together with text, so it can read screenshots, charts, and scanned documents. It returns text only; it does not generate images.
Can GPT-5.6 Luna use tools and JSON output?
Yes. It supports function calling and schema-constrained JSON responses, which makes it practical as a step inside automated workflows, data pipelines, and larger multi-model agent systems.
Should I use GPT-5.6 Luna or GPT-6 Luna for a new project?
OpenAI's model guidance marks GPT-5.6 as legacy and recommends the GPT-6 family for new work. Keep Luna 5.6 where an existing pipeline is tested on it and migrate when you next evaluate.
How much does GPT-5.6 Luna cost?
On TokenLab, GPT-5.6 Luna costs Input $0.06 / Output $0.36 per 1M tokens. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.
What are the context window and output limit of GPT-5.6 Luna?
GPT-5.6 Luna accepts up to 1,050,000 tokens of context and returns up to 128,000 tokens in one response.
Which endpoint should GPT-5.6 Luna use?
Use https://api.tokenlab.sh/v1/responses for GPT-5.6 Luna. The request example below shows the matching code shape.
Which operations does GPT-5.6 Luna support?
GPT-5.6 Luna supports Text to text. Select an operation above to see its endpoint and request example.
Compare GPT-5.6 Luna
Guides that use GPT-5.6 Luna
Sources
Reviewed Oct 2, 2026
More from OpenAI