Skip to main content
GET
Get test run status

Authorizations

X-API-Key
string
header
required

API key for authentication

Path Parameters

task_id
string
required

Test run to poll for status and results

Example:

"a3b2c1d0-e5f4-3210-abcd-ef1234567890"

Query Parameters

only_failed
boolean
default:false

Return only failing test cases. Omit to return every case

mode
enum<string>
default:full

How much of each test case to return. full returns every field of every case. summary returns one light row per case, with its ID, name, verdict and short reason, leaving out the conversation, the agent's output and the evaluator verdicts. Read those one case at a time from GET /agent-tests/run/{task_id}/results/{test_uuid}

Available options:
full,
summary
compact
boolean
default:false

Return a compact response that omits heavy detail fields (results.output, results.test_case, results.judge_results, results.reasoning, evaluators.output_config), keeping only the lightweight decision fields. Omit for full detail

Response

Successful Response

task_id
string
required

Test run job ID

Required string length: 36
Example:

"a3b2c1d0-e5f4-3210-abcd-ef1234567890"

name
string
required

Name of the run. A run nobody has renamed shows its number instead, such as Run 1 for a test run or Benchmark 1 for a benchmark

status
enum<string>
required

Current status of the run

Available options:
queued,
in_progress,
cancelled,
done,
failed
test_uuids
string[] | null

IDs of the tests this run executed, in run order

total_tests
integer | null

Total number of test cases

passed
integer | null

Number of test cases that passed

failed
integer | null

Number of test cases that did not pass, which includes the ones that produced no answer

latency_ms
object | null

Aggregated response latency in milliseconds, as {p50, p95, p99, count}

cost
object | null

Aggregated cost as {mean, min, max, count} (USD)

total_tokens
object | null

Aggregated token usage as {mean, min, max, count}

evaluators
TestRunEvaluator · object[] | null

The evaluators used in this run. Each verdict in judge_results links to one of these by evaluator_uuid

evaluator_summary
object[] | null

Totals for each evaluator over the whole run, matching the shape a benchmark reports for each model. Only evaluators that returned a verdict appear

results
TestCaseResult · object[] | null

Results for each test case

unanswered_tests
integer | null

Number of test cases that produced no answer because the agent or the judge could not be reached, which makes the pass rate an unfair measure of the agent

stopped_early
boolean
default:false

Whether the run stopped before starting every test case, after too many failed in a row

aborted
boolean
default:false

Whether a user stopped this run before it finished. The results collected up to that point are kept, and test cases that never ran are counted neither as passed nor as failed

error
string | null

Why the run could not be carried out, when it failed before producing any result

is_public
boolean
default:false

Whether the run is shared publicly

share_token
string | null

Token for building the public share URL