OrbitBench

A benchmark that scores LLMs on code generation.

Creator · May 2026 · Open source

Runs model-written code against hidden tests in isolation and reports a pass@1 score out of 100, from the CLI or a web dashboard.

Python