Test Result
Back to GlossaryResults
The outcome for one test in one test run — status, metric scores, and optional response text.
Also known as: results
Overview
A test result is a single row of execution data: one test evaluated in one Test Run. It stores pass/fail status, per-metric scores and reasoning, and optionally the model response text.
Do not confuse this with the Results Overview dashboard, which aggregates trends across many test runs.
What a test result includes
- Status — pass, fail, or error for that test
- Metric scores — only metrics whose includes the test's type and that are linked to the test's requirement
- Response — endpoint output captured during the run (when available)
- Links — , , requirement, category, topic
How results are produced
- A Test Set runs against an Endpoint → creates a test run
- Each test in the set executes (Single-Turn prompt or Multi-Turn conversation)
- Linked metrics evaluate the response → one test result per test
Querying results
- One run: with
- One test: by id
- Trends across runs: Results Overview UI or stats APIs — not a single test result entity
Related terms
- Test Run — parent execution snapshot
- Test — the case that was executed
- Metric — how the test was scored
- Metric Scope — determines which metrics run for this test's type