---
title: "WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation"
source: HuggingFace Daily Papers · 2026-05-17
url: https://arxiv.org/abs/2605.10912
date: 2026-05-18
published_at: 2026-05-17T12:00:00+00:00
tag: 论文研究
item_id: 9cd42b97cc1bf8fc
---
# Computer Science > Computation and Language

[Submitted on 11 May 2026]

# Title:WildClawBench: A Benchmark for Real-World, Long-Horizon Agent Evaluation

[View PDF](https://arxiv.org/pdf/2605.10912)

[HTML (experimental)](https://arxiv.org/html/2605.10912v1)

Abstract:Large language and vision-language models increasingly power agents that act on a user's behalf through command-line interface (CLI) harnesses. However, most agent benchmarks still rely on synthetic sandboxes, short-horizon tasks, mock-service APIs, and final-answer checks, leaving open whether agents can complete realistic long-horizon work in the runtimes where they are deployed. This work presents WildClawBench, a native-runtime benchmark of 60 human-authored, bilingual, multimodal tasks spanning six thematic categories. Each task averages roughly 8 minutes of wall-clock time and over 20 tool calls, and runs inside a reproducible Docker container hosting an actual CLI agent harness (OpenClaw, Claude Code, Codex, or Hermes Agent) with access to real tools rather than mock services. Grading is hybrid, combining deterministic rule-based checks, environment-state auditing of side effects, and an LLM/VLM judge for semantic verification. Across 19 frontier models, the best, Claude Opus 4.7, reaches only 62.2% overall under OpenClaw, while every other model stays below 60%, and switching harness alone shifts a single model by up to 18 points. These results show that long-horizon, native-runtime agent evaluation remains a far-from-resolved task for current frontier models. We release the tasks, code, and containerized tooling to support reproducible evaluation.

### References & Citations

Loading...

# Bibliographic and Citation Tools

Bibliographic Explorer

*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))
Connected Papers

*(*[What is Connected Papers?](https://www.connectedpapers.com/about))
Litmaps

*(*[What is Litmaps?](https://www.litmaps.co/))
scite Smart Citations

*(*[What are Smart Citations?](https://www.scite.ai/))# Code, Data and Media Associated with this Article

alphaXiv

*(*[What is alphaXiv?](https://alphaxiv.org/))
CatalyzeX Code Finder for Papers

*(*[What is CatalyzeX?](https://www.catalyzex.com))
DagsHub

*(*[What is DagsHub?](https://dagshub.com/))
Gotit.pub

*(*[What is GotitPub?](http://gotit.pub/faq))
Hugging Face

*(*[What is Huggingface?](https://huggingface.co/huggingface))
ScienceCast

*(*[What is ScienceCast?](https://sciencecast.org/welcome))# Demos

# Recommenders and Search Tools

Influence Flower

*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))
CORE Recommender

*(*[What is CORE?](https://core.ac.uk/services/recommender))# arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? [ Learn more about arXivLabs](https://info.arxiv.org/labs/index.html).