#code evaluation

AWS AI Labs has launched SWE-PolyBench, an open-source, multilingual benchmark designed to evaluate AI coding agents with real-world coding tasks across multiple languages, improving upon previous limited benchmarks.