Browse Benchmarks
Benchmark runs are not supported yet
In AI Inference, clicking New Benchmark Run does not create a run; the message "Creating benchmark runs is not available in AI Inference yet" appears. For now you can only view existing benchmark results.
Browse benchmark runs and view their results.
Step 1. Go to Benchmarks
Navigate to Evaluation → Benchmarks. The screen has two tabs:
- Dashboard: Highest/Average score, registered model count, score chart, and Model leaderboard summary.
- Benchmark Run: The run list (browse / create / delete). Select this tab to manage runs.
Benchmark Run List Columns
| Column | Description |
|---|---|
| Status | Run status |
| Name | Run name |
| Model | Model being evaluated |
| Benchmark | Benchmark type |
| Started at | Run start time |
| Action | Row menu (⊙ → View logs · View Results · Delete) |
Use New Benchmark Run (top right) to create a run, Refresh to reload, the search box to filter, and the Delete button to bulk-delete checked rows.
Step 2. View Results
Open a row's Action (⊙) menu:
| Menu | Description |
|---|---|
| View logs | View run logs |
| View Results | View benchmark scores/results |
| Delete | Delete the run |