challenges
The 13tasks every model runs. Full prompt, full rubric, no black box — click into any of them to see exactly what's being judged.
fizzbuzz
Classic FizzBuzz — tests basic correctness and code style
best so far: Claude Sonnet 4.6binary-search
Binary search implementation with full docs
best so far: Claude Haiku 4.5api-client
Write a small HTTP API client class with error handling and docs
best so far: Llama 3.3 70Breadme-writer
Write a README for a CLI tool — tests documentation ability directly
best so far: Claude Sonnet 4.6refactor
Refactor messy code and explain each change
best so far: Claude Haiku 4.5test-writing
Write pytest tests for a provided function — tests edge-case thinking and assertion quality
best so far: Claude Haiku 4.5debug
Find and fix 3 bugs in broken Python code — tests careful reading and correctness reasoning
best so far: Gemini 2.5 Flashasync-fetch
Write async Python for concurrent HTTP fetching with timeout and retry
best so far: Claude Sonnet 4.6sql
Write a complex SQL query with CTEs, window functions, and aggregations
best so far: Gemini 2.5 Flashgo-test
Write idiomatic Go table-driven tests — tests knowledge of Go testing conventions
best so far: Claude Haiku 4.5elixir-test
Write idiomatic Elixir ExUnit tests — tests knowledge of Elixir testing conventions
best so far: Claude Sonnet 4.6doom
Build a Doom-style raycasting FPS engine in a single HTML file — the mager-bench signature challenge
best so far: Claude Haiku 4.5slots
Build a Vegas-style slot machine in a single HTML file — reels, pay table, betting, wins
best so far: Claude Haiku 4.5