Supported Formats
Choose View Report in the Source column to open one of these four views. Select a format below to see an example.
| View | What it shows | Supported formats |
|---|---|---|
| Performance | Token speed, first-token latency, context and concurrency tests, with the reported hardware and settings. | |
| Evaluation | Task and scenario scores, scoring rules and any associated performance results. | |
| Sessions | Individual agent requests or test fixtures, with their timings, token counts and draft modes. | |
| Generation | Image and video results, including resolution, steps, batch size, cold/warm state and timing scope. |
Known formats use dedicated adapters. Other collected LLM results use a generic report. Serving tests, custom evaluations and coding sessions preserve their original measurements and settings. Viewer support does not verify a benchmark; original sources remain linked.