{"slug": "worldbench-culturally-grounded-benchmark-for-multilingual-agents", "title": "WorldBench: Culturally Grounded Benchmark for Multilingual Agents", "summary": "Researchers introduced WorldBench, a multilingual benchmark of 1,600 persona-grounded everyday tasks across seven languages and eight cultures, finding that frontier models achieve only 49.2% Constrained Task Success (CTS), revealing brittleness in multilingual agentic scenarios. The benchmark, detailed in an arXiv paper (2609.01056v1), extends prior metrics with CTS to evaluate task completion and environment preservation, showing large gaps between correctness and state preservation across models.", "body_md": "arXiv:2609.01056v1 Announce Type: new\nAbstract: 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", "url": "https://wpnews.pro/news/worldbench-culturally-grounded-benchmark-for-multilingual-agents", "canonical_source": "https://www.machinebrief.com/news/worldbench-culturally-grounded-benchmark-for-multilingual-ag-sr8k", "published_at": "2026-09-02 04:00:00+00:00", "updated_at": "2026-09-02 06:52:59.750738+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-research", "ai-agents"], "entities": ["WorldBench", "arXiv"], "alternates": {"html": "https://wpnews.pro/news/worldbench-culturally-grounded-benchmark-for-multilingual-agents", "markdown": "https://wpnews.pro/news/worldbench-culturally-grounded-benchmark-for-multilingual-agents.md", "text": "https://wpnews.pro/news/worldbench-culturally-grounded-benchmark-for-multilingual-agents.txt", "jsonld": "https://wpnews.pro/news/worldbench-culturally-grounded-benchmark-for-multilingual-agents.jsonld"}}