Docs / API
Models
GET /v1/models lists the active models with their four rates, context window and status.
Last updated
Endpoint
GET https://api.cent-ai.dev/v1/models
Authorization: Bearer cai_<secret>
The endpoint needs an API key. The same catalog is public on the pricing page of the website.
curl https://api.cent-ai.dev/v1/models \
-H "Authorization: Bearer $CENT_AI_KEY"
Response
{
"object": "list",
"data": [{
"id": "glm-5.3",
"object": "model",
"created": 1790000000,
"owned_by": "cent-ai",
"context_window": 1000000,
"pricing": {
"currency": "USD",
"per": "1M tokens",
"input": "0.070000000",
"cache_read": "0.013000000",
"cache_write": "0.070000000",
"cache_write_1h": null,
"output": "0.220000000",
"official_input": "1.400000000",
"official_output": "4.400000000",
"minimum_per_request": "0.000020000"
},
"pack_min_tier": 2,
"capabilities": ["tools", "reasoning"],
"status": "operational"
}]
}
The values above are an example of the shape, not current prices.
| Field | Meaning |
|---|---|
id | The model name to send in model. Names are stable |
context_window | Context size in tokens |
pricing.input, cache_read, cache_write, output | The four rates in USD per 1M tokens, see Pricing |
pricing.cache_write_1h | Rate for cache writes that live 1 hour; null when the model has no such rate |
pricing.official_input, official_output | The official reference price, used to show the actual discount |
pricing.minimum_per_request | Minimum charge per successful request |
pack_min_tier | The smallest budget pack that includes the model: 1 ($1.99), 2 ($9.99), 3 ($99); null for an image model |
pricing.variants | Image models only: a list of {name, unit, price, official_price}, the price per image (unit image) in USD. An image model is billed by these prices, not by the per-token fields; see Images |
status | operational, degraded, down or coming_soon |
Amounts are decimal strings with 9 fraction digits, so they never lose precision in a float.
Event models
During the event, some models are listed with a :free id and a price of 0. They have no billed id, so the name is the only one you can send:
| Id | Price | Pool |
|---|---|---|
hy4-flash:free | 0 for every rate | A fixed pool per UTC day; the current figure is on the event page |
These models are free for a limited time and may end at any time. When the day's pool is used up, the request answers 429 rate_limited with a Retry-After header until 00:00 UTC. The live remaining pool is on the event page.
Model names
Use the id exactly as listed. A name that does not exist, or a model that is not available, returns 404 model_not_found with the message "The model '{model}' does not exist or is not available."
Contract
GET/v1/models
List models with their rates
Auth · API key (Bearer or x-api-key)
Responses
- 200Active models.
- 401Error in the OpenAI shape.
Response headers
x-request-id
Generated from the OpenAPI contract.