Are Performance-Optimization Benchmarks Reliably Measuring Coding Agents?

A study by researchers at Singapore Management University questions the reliability of certain benchmarks used to evaluate coding agents, highlighting issues that may affect their performance scores.

1sec.ai