Skip to main content

Benchmark Scope & Categories

Verdict evaluates frontier models across 10 benchmark categories spanning interactive web apps, 2D game loops, vector artwork, and multi-step autonomous planning.

Interactive

Frontend UI

Landing pages, dashboards, responsive CSS, accessible forms.

Prompts Sourced: 10
View Category
Interactive

Game Dev

Browser 2D canvas games, physics loops, platformers, shooters.

Prompts Sourced: 10
View Category
Visual

SVG Art

Vector illustrations, generative math art, isometric cityscapes.

Prompts Sourced: 10
View Category
Autonomous

Agentic Tasks

Multi-step refactoring, execution planning, tool use.

Prompts Sourced: 10
View Category
Textual

Creative Writing

Technical documentation, release notes, lore generation.

Prompts Sourced: 10
View Category
Visual

3D Graphics

Three.js WebGL scenes, shader shaders, 3D meshes.

Prompts Sourced: 10
View Category
Interactive

Data Viz

Recharts, D3 SVG charts, heatmaps, interactive data graphs.

Prompts Sourced: 10
View Category
Interactive

Animation

CSS keyframe timelines, Framer Motion springs, micro-interactions.

Prompts Sourced: 10
View Category
Autonomous

Full-Stack

API routes, database ORM queries, auth middleware.

Prompts Sourced: 10
View Category
Algorithmic

Code Golf

Byte-minified solutions, algorithm optimization, math solvers.

Prompts Sourced: 10
View Category