GPT-5.6 Luna API
Efficient GPT-5.6 model for cost-sensitive agent and long-context workloads.
Inside the SU500 plan on the Azure channel, the landed discount is 8.8% off.
- Official input
- $0.6 / 1M
- Official output
- $3.6 / 1M
- Cache read
- $0.06 / 1M
- Cache write
- $0.75 / 1M
About GPT-5.6 Luna
Use for high-throughput GPT-5.6 workloads where cost efficiency matters more than peak capability.
- Context window
- 1050K tokens
- Max output
- 128K tokens
- Input
- text, image
- Tool calling
- Supported, parallel
- Web search
- Supported
- Prompt caching
- Supported
- Reasoning effort
- none / low / medium / high / xhigh / max
- API protocols
- chat.completions, responses
Pay-as-you-go discounts by channel
The same model can run through different channels at different landed discounts. Shorter bars cost less.
| Channel | Final discount | Quota multiplier | Billing unit |
|---|---|---|---|
| Azure Lowest | 5% off | 6.384× | USD / 1M |
Cache hits are billed at the cache read price. Cache writes use the cache write price, and the Console usage detail shows the final cost breakdown.
Plans that include this model
On a plan, calls are deducted from the daily quota at the plan channel's discount.
| Package | Final discount | Channel | Quota multiplier | Billing unit |
|---|---|---|---|---|
| SU500 | 8.8% off | Azure | 6.84× | USD / 1M |
| SU750 | 8.8% off | Azure | 6.84× | USD / 1M |
| SU1000 | 8.8% off | Azure | 6.84× | USD / 1M |
Example call
curl https://www.su8.codes/v1/chat/completions \ -H "Authorization: Bearer $SU8_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model": "gpt-5.6-luna", "messages": [{"role": "user", "content": "Hello"}]}'curl https://www.su8.codes/v1/chat/completions \ -H "Authorization: Bearer $SU8_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model": "gpt-5.6-luna", "messages": [{"role": "user", "content": "Hello"}]}' from openai import OpenAI client = OpenAI( base_url="https://www.su8.codes/v1", api_key="$SU8_API_KEY", ) client.chat.completions.create( model="gpt-5.6-luna", messages=[{"role": "user", "content": "Hello"}], ) # ~/.codex/config.toml model_provider = "su8" model = "gpt-5.6-luna" [model_providers.su8] name = "su8" base_url = "https://www.su8.codes/v1" wire_api = "responses"
Frequently asked questions
How much does the GPT-5.6 Luna API cost?
OpenAI list price: Input $0.6/1M, Output $3.6/1M, Cache read $0.06/1M. On SU8 Codes, each channel has its own landed discount; the current pay-as-you-go and package discounts are in the channel and package tables on this page, refreshed live.
How do I call GPT-5.6 Luna?
Create an API key in the SU8 Codes console, point your client at this site's base URL, and use the model ID gpt-5.6-luna. The example call above has copy-ready cURL and Python code.
Can I use GPT-5.6 Luna with Codex CLI?
Yes. In ~/.codex/config.toml, set the base URL to this site, add your SU8 API key, and set the model to gpt-5.6-luna. The Codex CLI tab in the example call above has the full config, and the integration docs walk through each step.
What is the context window of GPT-5.6 Luna?
1050K tokens. It also supports tool calling, prompt caching, reasoning effort.
Integration guides
More from OpenAI
| Model | Official input | Official output | Landed discount |
|---|---|---|---|
| GPT-5.6 Sol gpt-5.6-sol | $4/ 1M | $20/ 1M | 92.5% offCodex |
| GPT-5.6 Terra gpt-5.6-terra | $2/ 1M | $12/ 1M | 92.5% offCodex |
| GPT-5.5 gpt-5.5 | $5/ 1M | $30/ 1M | 92.5% offCodex |
| GPT-6-astra gpt-6-astra | $10/ 1M | $50/ 1M | 92.5% offCodex |
| GPT-5.4 Mini gpt-5.4-mini | $0.75/ 1M | $4.5/ 1M | 5% offAzure |
| gpt-image-2.6-sunburst gpt-image-2.6-sunburst | $0.15/ image | Per image | 92.5% offCodex |