Web Development
Interactive applications, visual tools, and layout or accessibility work verified in the browser. Includes the 26 multimodal tasks.
This category contains the 26 multimodal tasks, which require multimodal model capabilities during execution and are evaluated separately with seven supported models, plus one general task. See the multimodal leaderboard for results.
Interactive Applications
12 tasks
M001
Build a creator revenue dashboard from local campaign data
M002
Build a studio production task board
M003
Build an interactive pet care mini game
M004
Build an incident command console
M006
Build a local appointment finder UI
M007
Build an interactive budget planning workspace
M009
Build an interactive course builder app
M010
Build an ecommerce support ticket router
M012
Build a policy difference viewer
M013
Repair a broken Kanban build and preserve workflow behavior
M016
Build a feature flag administration panel
M020
Build a form builder with live validation
Visual Tools
11 tasks
M008
Build a local media timeline editor
M011
Build a local data explorer UI
M014
Retrofit a product UI for keyboard and semantic accessibility
M015
Redesign a product page for responsive use
M017
Build a travel itinerary planner
M018
Build scientific simulation controls
M021
Build a choropleth map filter over synthetic districts
M022
Build a waveform cue editor over a synthetic amplitude plot
M023
Build a local spreadsheet grid with protected cells and a visible formula bar
M024
Annotate a synthetic before/after image pair
M025
Explore a synthetic node-link graph with neighbor highlighting
Layout and Accessibility
4 tasks
M005
Build a research laboratory website
M019
Rebuild a portfolio from a reference style brief
M026
Build a paginated print preview over synthetic article pages
T031
Repair a Representative Accessibility Conformance Sample