LabyrinthBench: a local-focused, judge-free LLM benchmark that measures context recall under interference for multi-step agentic tasks.

LabyrinthBench measures the thing that actually kills long agent runs — whether a model can still use what it learned twenty turns ago — deterministically, with no LLM judge, on your own hardware, with a swappable harness for testing whatever context-management strategy you think will fix it.

中文全文翻译正在补齐，可先查看下方英文原文。

LabyrinthBench measures the thing that actually kills long agent runs — whether a model can still use what it learned twenty turns ago — deterministically, with no LLM judge, on your own hardware, with a swappable harness for testing whatever context-management strategy you think will fix it.