Model & Settings

Each model in the grid has its own independent settings. Click a model column header to adjust its configuration.

Settings

Temperature — controls randomness (0–2). Lower values produce more predictable output, higher values produce more creative output. Hidden for reasoning models.
Max Tokens — sets the maximum response length (1–128,000).
Reasoning Effort — for reasoning models only. Choose from Low, Medium, High, or Extra High.
Show Thinking — for reasoning models. Reveals the model's reasoning process in a collapsible section above the response.

Run tests

Run All — runs every cell in the grid. Progress shows in the toolbar.
Run a row — hover a version label and click the play icon to test one version across all models.
Run a column — hover a model header and click the play icon to test all versions on one model.
Run one cell — double-click any cell or use its play icon.

If your prompt uses template variables, fill them in on the Setup panel before running. Your values persist between runs.