Short answer
GPT-5.6 Luna is available through APIVAI at $0.0304 per million input tokens and $0.18 per million output tokens as of 2026-10-07, 85% below OpenAI's official list price of $0.20 / $1.20. Use the model ID gpt-5.6-luna with an APIVAI key in the OpenAI format at https://api.apivai.com/v1 (Chat Completions and the Responses API). Billing is pay-as-you-go from a prepaid balance, from $10, with no subscription.
GPT-5.6 Luna API pricing
| Per 1M tokens | APIVAI | Official list price |
|---|---|---|
| Input | $0.0304 | $0.20 |
| Output | $0.18 | $1.20 |
Prompt caching: cache writes cost $0.0384 and cache reads $0.0032 per million tokens. Context window: 1M tokens. Prices are checked against the price list every 6 hours; see all model prices.
How much does GPT-5.6 Luna cost per request?
Each request is billed for its input and output tokens. Coding agents resend the conversation and files with every step, so their requests are much larger than a chat message.
| Example | Input / output tokens | APIVAI | Official price |
|---|---|---|---|
| One chat message | 2,000 / 500 each | $0.000151 | $0.001 |
| One coding-agent step | 30,000 / 1,500 each | $0.00118 | $0.0078 |
| 1,000 chat messages | 2,000 / 500 each | $0.151 | $1.00 |
| A month of agent use (100 steps a day, 22 days) | 30,000 / 1,500 each | $2.60 | $17.16 |
Thinking is on by default and thinking tokens are billed as output, so the real output count can be higher than the text you see.
How do you call GPT-5.6 Luna?
cURL (OpenAI format)
curl https://api.apivai.com/v1/chat/completions \
-H "Authorization: Bearer $APIVAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-5.6-luna",
"max_tokens": 4096,
"messages": [{"role": "user", "content": "Hello!"}]
}'Python (OpenAI SDK)
from openai import OpenAI
client = OpenAI(api_key="YOUR_APIVAI_API_KEY", base_url="https://api.apivai.com/v1")
resp = client.chat.completions.create(
model="gpt-5.6-luna",
max_tokens=4096,
messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)Codex CLI
# ~/.codex/config.toml model = "gpt-5.6-luna" model_provider = "apivai" [model_providers.apivai] name = "APIVAI" base_url = "https://api.apivai.com/v1" env_key = "OPENAI_API_KEY" wire_api = "responses"
Codex uses the Responses API, which APIVAI supports. More tools: API documentation and OpenAI-compatible API.
Other OpenAI models
| Model | Input / output per 1M | Below official price |
|---|---|---|
| GPT-6 Astra | $1.52 / $7.60 | 85% |
| GPT-5.5 | $0.77 / $4.56 | 85% |
| GPT-5.6 Sol | $0.61 / $3.65 | 85% |
| GPT-5.6 Terra | $0.30 / $1.82 | 85% |
| GPT-6.1 Sol | $0.30 / $1.52 | 85% |
| GPT-6 Sol | $0.30 / $1.52 | 85% |
| GPT-6 Luna | $0.0144 / $0.0768 | 86% |
Setup guides
FAQ
Is this the same GPT-5.6 Luna model?
Yes. Requests with the model ID gpt-5.6-luna are answered by GPT-5.6 Luna; only the price and the way you pay differ.
What is the model ID for GPT-5.6 Luna?
gpt-5.6-luna. The list of models changes over time, so check GET https://api.apivai.com/v1/models with your key before hard-coding a name.
Which tools can use GPT-5.6 Luna?
Codex CLI and any OpenAI-compatible tool or SDK (Cursor, Cline, Continue, Aider and others) through https://api.apivai.com/v1.
How is GPT-5.6 Luna billed?
Per token, from a prepaid balance: $0.0304 per million input tokens and $0.18 per million output tokens, with thinking tokens counted as output. The minimum top-up is $10 and there is no subscription.
How many requests can I send?
Each key allows 60 requests per minute by default; support can raise the limit.
Create an account, top up from $10 and use gpt-5.6-luna with your key.