ToolMaze Benchmark

What is ToolMaze?

A benchmark that tests how tool-using LLM agents replan and recover from tool failures. It varies task structure (DAG-based complexity) and failure type…

Released

2026-06-04

Evaluates

General AI, Language & Knowledge, cs.AI

Openness

unknown

Importer

Claire Radar

Review status

ai-reviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2606.05806

ToolMaze paper and code

• Paper https://arxiv.org/abs/2606.05806
• Code https://github.com/Zhudongsheng75/ToolMaze

Which sources cite ToolMaze?

• ToolMaze

Related General AI, Language & Knowledge, cs.AI benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• A3S-Bench
• AARRI-Bench
• ABC-Bench
• ActBench