ToolFailBench
Explore score history ↗ToolFailBench evaluates tool-use failures in LLM agents across 1,000 tasks in finance, medicine, law, cybersecurity, and real estate. It labels traces with Tool-Skip, Result-Ignore, Output-Fabrication, and Unnecessary-Tool-Use, using a rule classifier and LLM judges.
- Released
- 2026-07-06
- Institutions
- University of California, Berkeley
- Evaluates
- Cybersecurity, Finance, Language & Knowledge, cs.CL
- Openness
- unknown
- Importer
- Claire Radar
- Review status
- unreviewed
ToolFailBench paper
No reported scores are on record for this benchmark yet.