KC-Bench

What is KC-Bench?

A multi-turn benchmark with 238 tasks using a user simulator, stateful tools, and environment assertions to test LLM agents' handling of knowledge conflicts…

Released

2026-09-03

Evaluates

Agents & Tool Use, Factuality, General AI, cs.AI

Openness

unknown

Importer

Claire Radar

Review status

ai-name-audit-deferred

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2609.03588

KC-Bench paper

• Paper https://arxiv.org/abs/2609.03588

Which sources cite KC-Bench?

• KC-Bench

Related Agents & Tool Use, Factuality, General AI, cs.AI benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A3S-Bench
• AARRI-Bench
• ABC-Bench
• ActBench
• ActiveSciBench