Metraipolis › archive › dc_f575166018ff

WorldBench: Culturally Grounded Benchmark for Multilingual Agents — arXiv:2609.01056

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.01056
License (as archived): CC0-1.0
Archived: 2026-10-17T06:32:19.238Z
SHA-256 of the archived text: 1407eb5f6199c801e68d24a9e7cd60e3b929ad9ad78c21e3ce77eebceb34151a
Stable URL of this page: https://metraipolis.com/documento/dc_f575166018ff

How to cite

WorldBench: Culturally Grounded Benchmark for Multilingual Agents — arXiv:2609.01056 — arXiv (en), https://arxiv.org/abs/2609.01056 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f575166018ff (Metraipolis archive dc_f575166018ff, sha256 1407eb5f6199c801e68d24a9e7cd60e3b929ad9ad78c21e3ce77eebceb34151a)

BibTeX

@misc{dcf575166018ff,
  title = {WorldBench: Culturally Grounded Benchmark for Multilingual Agents — arXiv:2609.01056},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.01056},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f575166018ff; sha256 1407eb5f6199c801e68d24a9e7cd60e3b929ad9ad78c21e3ce77eebceb34151a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.01056. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Despite the growing use of LLM-powered agents to solve multi-step tasks in complex environments, existing benchmarks rarely test state preservation, performance across languages, and application to realistic, grounded scenarios. To address these concerns, we present WorldBench: a comprehensive, multilingual benchmark of genuine, persona-grounded everyday workflows, where agents can act in a sandbox via structured actions. WorldBench comprises 1,600 tasks across seven languages and eight cultures, filtered and refined through feedback from human annotators with language- and culture-specific expertise. For evaluation, we extend metrics from previous works and introduce Constrained Task Success (CTS), which combines natural language instructions and testbeds to score task completion, minimal modification, and other complementary metrics through deterministic and LLM-as-a-Judge evaluations. Our experiments show that frontier models reach only 49.2% CTS, with all models demonstrating large gaps between correctness and environment preservation. We thereby show that current agents remain brittle in multilingual, agentic scenarios, especially for long-horizon tasks and under state-preservation constraints

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.01056 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-17T06:15:12.672Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).