API reference

API reference · API

Cancel a run

POST/v1/evals/{id}/runs/{runId}/cancelAPI key

Chat scope. Nothing pending starts; a case in flight is graded and paid for as it would have been. Models graded on every case before the cancel still leave evidence.

Parameters

Path and query parameters
FieldTypeWhat it is
id (path)requiredstringThe suite id (ev_...).
runId (path)requiredstringThe run id (evrun_...).

Answer 200

The run.

Fields of the 200 answer
FieldTypeWhat it is
idstring
object"eval.run"
suite_idstring
statusstringOne of "queued", "running", "completed", "budget_exhausted", "cancelled".
triggerstringOne of "manual", "scheduled".
revisioninteger
modelsarray of string
case_countinteger
budget_microsinteger
spent_microsinteger
reserved_microsintegerWorst cases held for cases in flight.
progressobject
progress.doneinteger
progress.totalinteger
created_atinteger
started_atinteger or null
finished_atinteger or null
thresholdnumber
cheapest_passingstring or nullThe cheapest model graded on every case at or over the threshold.
summaryarray of object
summary[].modelstring
summary[].pass_ratenumber or nullWeighted share of graded cases passed.
summary[].passedinteger
summary[].failedinteger
summary[].errorsinteger
summary[].skippedinteger
summary[].pendinginteger
summary[].gradedinteger
summary[].casesinteger
summary[].completeboolean
summary[].cost_microsinteger
summary[].cost_per_case_microsinteger or null
summary[].median_latency_msinteger or null
summary[].on_frontierbooleanNo other model is as cheap or cheaper and passes as much or more.
failuresarray of objectAt most 200, each with the model's output clipped to 4000 characters.
failure_countinteger

Errors

  • 400OpenAI shaped error.
  • 401OpenAI shaped error.
  • 404OpenAI shaped error.

Every error answers in one shape, with its type, code and the field it names. The whole list is in the docs' Errors section.

Antimatter