Active-SWE Benchmark

What is Active-SWE?

Active-SWE evaluates coding agents on proactive bug fixing: detecting and fixing multiple bugs without issue reports. It includes 1,663 tasks across six bug…

Released

2026-08-05

Evaluates

Code & Software, Software & AI Compute, cs.SE

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2608.04682

Active-SWE paper and code

• Paper https://arxiv.org/abs/2608.04682
• Code https://github.com/XLearning-SCU/Active-SWE

Which sources cite Active-SWE?

• Active-SWE

Related Code & Software, Software & AI Compute, cs.SE benchmarks

• AFT-Bench
• Agent App Benchmark
• Agent Retrieval Bench
• Agent Systems Benchmark (ASB)
• Agentic Evaluation in 2D Mazes
• Agentic Security Benchmark
• AgentInfraBench
• AgentSLABench