ExtractBench: A Benchmark for Schema-Guided Enterprise Document Extraction

  • 类型:arxiv
  • 标识:2607.29677
  • 链接:https://arxiv.org/abs/2607.29677
  • 主分类:evaluation
  • 形态:benchmark
  • 被引:1
  • 被引来源:Semantic Scholar
  • S2被引:1
  • OpenAlex被引:0
  • 影响力被引:0
  • TLDR:LlamaExtract Agentic Plus ranks first on all three metrics, with accuracy comparable to coding agents at a fraction of the cost, and is the first to score value accuracy, record completeness at scale, grounding, and measured cost together.
  • OpenAlex ID:W7196924744
  • OpenAlex DOI:10.48550/arxiv.2607.29677
  • DOI:10.48550/arxiv.2607.29677
  • DOI来源:OpenAlex
  • 开放获取:green
  • 开放获取链接:https://doi.org/10.48550/arxiv.2607.29677
  • OpenAlex更新:2026-08-26
  • 副分类:agent
  • 待LLM分类:否
  • 标题中文:ExtractBench:一个面向模式引导的企业文档抽取基准
  • TLDR中文:LlamaExtract Agentic Plus在三项指标上均排名第一,准确度可与coding agent相媲美而成本仅为其一小部分,是首个同时在大规模下对数值准确性、记录完整性、grounding与实测成本进行打分的方法。
  • 来源文件
  • /inbox/tom/_candidates/2026-08-03-agent-rag-longcontext-candidates.json
  • [S2 enrich]
  • [OpenAlex backfill]