hatchmoment. scored by care · not by stars

CodexProLong

Open-source toolkit solving EinsteinArena verification benchmarks

The project tackles the challenge of solving formal verification benchmarks from EinsteinArena, offering concrete implementations for each problem. It includes source code, artifacts, and replay receipts that prove the correctness of each solution. Researchers and developers in formal methods can use it to benchmark their own tools or to study verified constructions. Unlike ad‑hoc scripts, it provides a reproducible, documented pipeline with CI integration.

autonomous-agentscodexeinstein-arenamathematical-optimizationprogram-synthesisreproducible-researchscientific-computing
View on GitHub →

JamesWeatherhead/CodexProLong