LLMReasoningCapability

Whether a model exposes an OpenAI-style reasoning_effort control.

supportedSupported (boolean)

기본값: false

levelsstring[]

Possible values: [low, medium, high, xhigh]

default object

하나 이상(anyOf)

• string
• null

string

Possible values: [low, medium, high, xhigh]

LLMReasoningCapability

{
  "supported": false,
  "levels": [
    "low"
  ],
  "default": "low"
}