LLMDetectedLimits

Attributes the endpoint reports for the requested model — token limits plus (Anthropic) reasoning-effort capability.

context_window object

detected context window (tokens)

anyOf

• integer
• null

integer

max_tokens object

detected max output tokens cap

anyOf

• integer
• null

integer

reasoning_levels object

Supported reasoning-effort levels advertised by the endpoint, or null when the endpoint reports no reasoning capability.

anyOf

• string[]
• null

Array [

string

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

]

LLMDetectedLimits

{
  "context_window": 0,
  "max_tokens": 0,
  "reasoning_levels": [
    "low"
  ]
}