Aider Code Editing Benchmark

Best of 2024 in Agents (from #1 on SWE-Bench Full, Prof. Graham Neubig of OpenHands/AllHands)

Latent Space

A benchmark for evaluating agents on individual file code editing tasks.