AgentAbstain Benchmark

What is AgentAbstain?

AgentAbstain is a paired-task benchmark for evaluating LLM agents' ability to abstain from acting in scenarios such as ambiguity, conflicting constraints, or…

Released

2026-07-11

Evaluates

Agents, General AI, Language & Knowledge, Reasoning

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2607.10059

AgentAbstain paper

• Paper https://arxiv.org/abs/2607.10059

Which sources cite AgentAbstain?

• AgentAbstain

Related Agents, General AI, Language & Knowledge, Reasoning benchmarks

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