GPT-5.6 Luna
gpt-5.6-luna
gpt-5.6-luna is available at gpt-5.6-luna. Below: prices per channel, specifications and request examples.
Prices per channel
| Channel | Input, per 1M | Output, per 1M | Cache, per 1M |
|---|---|---|---|
| Auto · default gpt-5.6-luna | $0.0033 | $0.0195 | $0.0004 |
| Channel 2 gpt-5.6-luna:c2 | $0.0033 | $0.0195 | $0.0004 |
| Channel 5 gpt-5.6-luna:c5 | $0.0143 | $0.0856 | $0.0015 |
Prices per 1M tokens. You pay for what you use. Pick a channel with a suffix: gpt-5.6-luna:c2; without one you get the default channel.
Specifications
- Context window
- 1,050,000 tokens
- Max output tokens
- 128,000 tokens
- Image input
- yes
- Reasoning
- yes
- Anthropic format (/v1/messages)
- yes
- Data in Russia
- no
Request example
Create a key in the dashboard. Streaming works too: add "stream": true.
curl https://demo-en.alexshkarupa.com/v1/chat/completions \
-H "Authorization: Bearer $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://demo-en.alexshkarupa.com/v1", api_key="YOUR_KEY")
reply = client.chat.completions.create(
model="gpt-5.6-luna",
messages=[{"role": "user", "content": "Hello!"}],
)
print(reply.choices[0].message.content)Questions and answers
How much does gpt-5.6-luna cost?
Input is $0.0033 per 1M tokens, output $0.0195 per 1M. You are charged for the tokens actually used, so a single request costs a fraction of a cent. Your balance and spending are in the dashboard.
What is the context window of gpt-5.6-luna?
Up to 1,050,000 input tokens and up to 128,000 output tokens.
How do I call gpt-5.6-luna?
Set the base URL to https://demo-en.alexshkarupa.com/v1 and use "gpt-5.6-luna" in the model field of any OpenAI-compatible client. curl and Python examples are on this page.
How do channels differ?
Channels are independent routes to the same model at different prices. Without a suffix a request takes the default channel; to pin a channel add its number: gpt-5.6-luna:c2. A pinned channel never fails over silently: if it is unavailable, the request returns an error.
Is cached input cheaper?
Repeated context is billed at the cache price — $0.0004 per 1M tokens instead of $0.0033.
Does gpt-5.6-luna understand images?
Yes, images can be sent along with text.
Does it support reasoning?
Yes, the model supports a reasoning mode; its depth is set by a request parameter.
Other models
Start using gpt-5.6-luna in minutes
Sign up by email, create a key in the dashboard, top up your balance without a contract.