Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agentic Evaluation

Agent Agent Eval Benchmarks
在日益增多的智能体(agentic)基准测试上评估智能体极具挑战性,因为这些基准往往需要构建复杂的运行环境,并涉及与各类智能体系统的深度集成。为此,我们提出“Harbor 适配器”(Harbor Adapters)——一套面向智能体基准测试的统一化评估基础设施。本工作主要包含以下三方面贡献: 第一,我们开发了基准适配器(benchmark adapters),将逾 80 个现有基准测试统一迁移到可评估任意智能体的标准化框架中,并通过严格的代码审查与等价性实验(parity experiments)对适配器的正确性与一致性进行了充分验证; 第二,我们在涵盖不同能力层级的 8 个模型上,开展了覆盖 54 个基准测试的大规模系统性评估;每个模型均分别采用 Terminus-2 框架及三种原生评测套件(native harnesses)之一进行测试。该设计显著拓展了对智能体能力边界与典型失效模式的分析广度与深度,远超以往研究的覆盖范围; 第三,我们提出了“Harbor 指数”(Harbor-Index)——一个经过精心筛选、由 82 项高难度、高多样性、高质量任务构成的精简评测集,其任务覆盖 29 个基准测试,系从上述适配后的完整基准套件中,经难度筛选、AI 辅助审核与人工复核双重把关,并辅以“审核—修正”迭代优化流程(audit-and-fix loop)提炼而成。Harbor 指数在保留大规模智能体评估所必需的挑战性与覆盖面的同时,大幅降低了计算成本,具备良好的可执行性:所有被评测的模型—套件组合均未超过 30% 的任务通过率,其中表现最强的配置(GPT-5.5 搭配 Codex)也仅达到 28.0%。我们已将全部适配器、评测结果、深入分析报告以及 Harbor 指数作为开源资源全面公开,旨在为语言模型智能体的评估提供更可靠、更全面的技术支撑。
Evaluating agents on the growing number of agentic benchmarks is challenging because they often require complex environments and agent integrations. We introduce Harbor Adapters, a unified evaluation infrastructure for agentic benchmarks. Our work makes three contributions. First, we develop benchmark adapters that port more than 80 benchmarks to evaluate arbitrary agents, and validate them through rigorous code review and parity experiments. Second, we conduct a large-scale evaluation of 8 models spanning capability tiers across 54 benchmarks; every model is run with Terminus-2 and with one of 3 native harnesses. This enables a broader analysis of agent capabilities and failure modes than was previously possible. Third, we introduce Harbor-Index, a curated set of 82 difficult, diverse, and high-quality tasks spanning 29 benchmarks, refined from the adapted suite through difficulty filtering, AI and human audit, and an audit-and-fix loop. Harbor-Index preserves the challenge and breadth of large-scale agentic evaluations while being affordable to run; no evaluated model-harness configuration exceeds 30% pass rate, and the strongest (GPT-5.5 with Codex) reaches 28.0%. We release the adapters, evaluation results, in-depth analysis, and Harbor-Index as open-source artifacts to support more reliable and comprehensive evaluation of language-model agents.
许愿