Skip to main content
Runs represent a single execution of a Benchmark. When you create a run, Verial provisions a playground, executes each task, evaluates the results, and produces a score and verdict. Each run contains Task Runs with individual results.

​
Endpoints

MethodEndpointDescription
GET/runsList runs
POST/runsCreate a run
GET/runs/{id}Get run details
POST/runs/{id}/completeMark a run as complete
POST/runs/{id}/cancelCancel a run

​
Run Object

FieldTypeDescription
idstringUnique identifier
benchmark_idstringParent Benchmark
statusStatusCurrent status (pending, running, completed, cancelled, failed)
scorenumber | nullOverall score (0-1), set on completion
verdictVerdict | nullPass/fail verdict, set on completion
organization_idstringParent organization
created_atdatetimeCreation timestamp
updated_atdatetimeLast modification timestamp
completed_atdatetime | nullWhen the run finished

​
SDK Example

// Create a run
const run = await verial.runs.create({
  benchmarkId: 'bench_abc123',
})

// List all runs
const runs = await verial.runs.list()

// Get run details (includes score and verdict)
const details = await verial.runs.get({ id: run.id })

// Mark as complete
await verial.runs.complete({ id: run.id })

// Cancel a running benchmark
await verial.runs.cancel({ id: run.id })