AgentSWE Benchmark

What is AgentSWE?

As the production of software passes into the hands of software itself, coding agents are increasingly commissioned to build agents fitted to their users' own…

Released

2026-10-09

Evaluates

Agents, Code, Code & Software, Software & AI Compute

Openness

unknown

Importer

Claire Radar

Review status

rule-auto-admitted

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2610.13361

AgentSWE paper and code

• Paper https://arxiv.org/abs/2610.13361
• Code https://github.com/VectorSpaceLab/AgentSWE

Which sources cite AgentSWE?

• AgentSWE

Related Agents, Code, Code & Software, Software & AI Compute benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• A3S-Bench
• Active-SWE
• Adaptive Adversaries
• AeroCopilotBench
• AFT-Bench
• Agent App Benchmark
• Agent Failure Lab