Models
The catalog, generated from the same pinned sheet the meter bills from. The model id is what you pass as model; context window and max output are in tokens. Rates are on the price page.
| Model id | Name | Context window | Max output tokens |
|---|---|---|---|
glm-5.3 | GLM-5.3 | 1,000,000 | 131,072 |
glm-5.2 | GLM-5.2 | 1,000,000 | 131,072 |
glm-5.1 | GLM-5.1 | 202,752 | 32,768 |
kimi-k3 | Kimi K3 | 1,048,576 | 131,072 |
kimi-k2.7-code | Kimi K2.7 Code | 262,144 | 262,144 |
kimi-k2.6 | Kimi K2.6 | 262,144 | 65,536 |
mimo-v2.5-pro | MiMo V2.5 Pro | 1,048,576 | 128,000 |
mimo-v2.5 | MiMo V2.5 | 1,000,000 | 128,000 |
minimax-m3 | MiniMax M3 | 1,000,000 | 131,072 |
minimax-m2.7 | MiniMax M2.7 | 204,800 | 131,072 |
qwen3.8-max | Qwen3.8 Max | 1,000,000 | 131,072 |
qwen3.7-plus | Qwen3.7 Plus | 1,000,000 | 65,536 |
qwen3.7-max | Qwen3.7 Max | 1,000,000 | 65,536 |
qwen3.6-plus | Qwen3.6 Plus | 1,000,000 | 65,536 |
hy3 | Hy3 | 256,000 | 64,000 |
longcat-2.0 | LongCat 2.0 | 1,000,000 | 131,072 |
muse-spark-1.2 | Muse Spark 1.2 | 1,048,576 | 131,072 |
deepseek-v4-pro | DeepSeek V4 Pro | 1,000,000 | 384,000 |
deepseek-v4-flash-vision-exp | DeepSeek V4 Flash Vision Exp | 1,000,000 | 384,000 |
deepseek-v4-flash | DeepSeek V4 Flash | 1,000,000 | 384,000 |
grok-4.6 | Grok 4.6 | 500,000 | 500,000 |
gpt-5.6-luna | GPT 5.6 Luna | 1,050,000 | 128,000 |
Notes#
- Every model works on all three wires. Pass any id to
/v1/chat/completions,/v1/responsesor/v1/messages; no model is restricted to one endpoint. max_tokensabove the max output is lowered, not rejected. Ask for more than a model serves and the request goes through at the model's cap. See Limitations for one model where the cap does not hold.- Context windows are enforced by the model, not by Sator. A prompt longer than the window is rejected as a
400or422with the model's own message relayed. See Errors. - Four models are tiered.
qwen3.7-plus,qwen3.6-plus,grok-4.6andgpt-5.6-lunaswitch to a second rate set for the whole request when the input side exceeds 200,000 tokens. Both rate sets are on the price page. - Discover them programmatically with
GET /v1/models, which returns exactly these ids and needs no key.
Models may be added, changed or removed at any time; this page and GET /v1/models are regenerated from the sheet on every deploy, so they never disagree with each other or with your bill.