EvalCaseResult
The result of a single EvalCase within an EvalRun.
Shape: object
| Property | Required | Type | Description | Rules |
|---|---|---|---|---|
id |
no | string (uuid) |
The result record identifier. | — |
evalRunId |
yes | string (uuid) |
The identifier of the run this result belongs to. | — |
caseId |
yes | string (uuid) |
The identifier of the case being measured. | — |
runId |
no | string (uuid) |
The identifier of the run record created while processing this case. This lets an eval failure jump straight to its transcript and span tree. | — |
passed |
yes | boolean |
Reports whether the case passed all of its checks. | — |
output |
no | string |
The text output produced by the agent. | — |
scores |
no | JsonElement |
Per-check score list (free-form JSON). | — |
failureReason |
no | string |
The failure reason. Populated only when Passed is false. |
— |