Skip to main content
Per-kind counts (HTTP 400 outside these ranges):

​
Scale is a fixed rubric

Always send all five levels 0 through 4. No sparse rubrics, no alternate ranges, no fewer or more levels. Each level must be a distinct integer in that set; anything else is rejected with HTTP 400. See Scale decisions for the exact shape.

​
Content and question length

The full rendered request must fit the plan’s context ceiling: roughly 32K tokens by default, or up to 128K tokens on Growth (Extended context). That budget includes:
  • content
  • question.instructions
  • option descriptions (Choice)
  • scale rubric lines (Scale)
  • sort item bodies (Sort)
  • tag specs (Tags)
  • any grounding-added search context
Oversized requests fail rather than silently truncating. Very long content can also hit HTTP 503 before the hard token cap, so keep documents well under the 32K ceiling in practice. See Errors for status codes and example error bodies.