▲ 2 pointsSlopCodeBench: Benchmarking How Coding Agents Degrade over Long-Horizon Tasksscbench.aiby matt_d·4mo ago·0 comments·view on hn ↗