Exact prompt sent to every model
Build a single-file HTML page containing a reusable data table component in vanilla JS (no libraries). Seed it with the 24 invoice rows below. Requirements:
- Columns: Invoice #, Client, Issued, Due, Amount, Status
- Click a header to sort ascending, again for descending, with a visible sort indicator; sorting must be type-aware (dates as dates, amounts as numbers)
- Text filter box that filters across Client and Invoice # as you type
- Status filter chips: All, Paid, Overdue, Pending (multi-select)
- Row checkboxes plus a header checkbox with indeterminate state; a toolbar shows "N selected" and a "Mark paid" action that updates the Status of selected rows
- Overdue rows have a subtle visual treatment; amounts right-aligned and formatted as USD
- Empty state when filters match nothing
- Dense but readable. Sticky header row. Works at 390px by allowing horizontal scroll of the table only. Return the complete file in one code block.
Responses
Latest published run: gemini-3.8-flash. For each model the repeat closest to its published median is shown; every repeat is in the repository.
No published attempts for this task yet
When the run is published, every model's raw response, per-dimension scores, judge rationales, human review notes and rendered artifacts appear here side by side.