OmnilingualGAIA2: Evaluating the Multilingual Gap in Frontier AI Agents

• URL: http://arxiv.org/abs/2608.08775v2
• Date: Tue, 11 Aug 2026 11:49:45 GMT
• Title: OmnilingualGAIA2: Evaluating the Multilingual Gap in Frontier AI Agents
• Abstract summary: We introduce OmnilingualGAIA2, a machine-translated expansion of the GAIA2 agentic benchmark, covering ten target languages spanning five writing systems.<n>We find a universal cross-lingual gap of 8.8-18.4 pass@3 points that is agent-asymmetric in magnitude, concentrates on tool-orchestration rather than quantitative reasoning, and does not close with model scale.<n>Human-expert linguistic analysis further identifies morphological cue loss and amplified ambiguity as the primary failure mechanisms in non-Latin-script languages.
• Score: 32.227608559573525
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Agentic benchmarks aim to measure how well AI agents plan, search, execute, and recover within realistic multi-tool environments, but they are almost exclusively in English. As AI agents are globally deployed to a linguistically diverse user base, whether agentic competence measured in English transfers to other languages remains an open question. We introduce OmnilingualGAIA2, a machine-translated expansion (with partial human- expert validation) of the GAIA2 agentic benchmark, covering ten target languages spanning five writing systems, paired with a localised and human-calibrated multilingual verifier. Evaluating seven frontier and open-weight agents, we find a universal cross-lingual gap of 8.8-18.4 pass@3 points that is agent-asymmetric in magnitude, concentrates on tool-orchestration rather than quantitative reasoning, and does not close with model scale. A stratified error attribution decomposes the gap as predominantly model-driven (55%), with a bounded translation-contamination floor of only 6.4% of scenario-language pairs. Human-expert linguistic analysis further identifies morphological cue loss and amplified ambiguity as the primary failure mechanisms in non-Latin-script languages. Our results argue that multilingual agentic evaluation must become a standard part of the reporting protocol for globally deployed agents.

Related papers

• Beyond Cross-Lingual Transfer: Benchmarking Propagation Boundaries in Multilingual LLM Unlearning [80.05116806217397]
  Large Language Model (LLM) unlearning aims to suppress target knowledge while preserving general capabilities.<n>We introduce CLLPU (Cross-Lingual and Language-Bound Protocol for LLM Unlearning), a multilingual benchmark that formulates this problem through two settings.<n>Common-goal forgetting, where target knowledge should be suppressed across all languages, and language-conditioned forgetting, where suppression should remain confined to a designated language.
  arXiv  Detail & Related papers  (2026-09-05T08:26:15Z)
• Multilingual Refusal Alignment for Safer Large Language Models [53.64286756804503]
  We investigate whether single-language alignment transfers cross-lingually, how language consistency is preserved during training, and the resulting trade-offs with general knowledge capabilities.<n>We introduce RefusEU, a novel refusal alignment dataset covering 12 European languages, including a dedicated test set for evaluating current state-of-the-art models.<n>Our controlled Direct Preference Optimization (DPO) experiments provide two key insights: aligning models exclusively in English is insufficient to ensure cross-lingual safety, even for the same harm categories, whereas training on multilingual datasets can improve safety without degrading general performance, as measured by the Global MMLU benchmark.
  arXiv  Detail & Related papers  (2026-04-24T09:29:14Z)
• Litmus (Re)Agent: A Benchmark and Agentic System for Predictive Evaluation of Multilingual Models [18.040787413825793]
  We introduce a benchmark of 1,500 questions spanning six tasks and five evidence scenarios.<n>The benchmark separates accessible evidence from ground truth, enabling evaluation of systems that must infer missing results from incomplete literature evidence.<n>We also present Litmus (Re)Agent, a DAG-orchestrated agentic system that decomposes queries into hypotheses, retrieves evidence, and synthesises predictions through feature-aware aggregation.
  arXiv  Detail & Related papers  (2026-04-10T05:16:33Z)
• Parallel Universes, Parallel Languages: A Comprehensive Study on LLM-based Multilingual Counterfactual Example Generation [49.2073409243885]
  Large language models (LLMs) excel at generating English counterfactuals and demonstrate multilingual proficiency.<n>We conduct automatic evaluations on both directly generated counterfactuals in the target languages and those derived via English translation across six languages.<n>We identify and categorize four main types of errors that consistently appear in the generated counterfactuals across languages.
  arXiv  Detail & Related papers  (2026-01-01T08:53:49Z)
• MuBench: Assessment of Multilingual Capabilities of Large Language Models Across 61 Languages [33.450081592217074]
  We introduce MuBench, a benchmark covering 61 languages and evaluating a broad range of capabilities.<n>We evaluate several state-of-the-art multilingual LLMs and find notable gaps between claimed and actual language coverage.
  arXiv  Detail & Related papers  (2025-06-24T09:53:00Z)
• Mechanistic Understanding and Mitigation of Language Confusion in English-Centric Large Language Models [56.61984030508691]
  We present the first mechanistic interpretability study of language confusion.<n>We show that confusion points (CPs) are central to this phenomenon.<n>We show that editing a small set of critical neurons, identified via comparative analysis with a multilingual-tuned counterpart, substantially mitigates confusion.
  arXiv  Detail & Related papers  (2025-05-22T11:29:17Z)
• MAPS: A Multilingual Benchmark for Global Agent Performance and Security [8.275240552134338]
  We propose MAPS, a benchmark suite designed to evaluate agentic AI systems across diverse languages and tasks.<n>We translate each dataset into eleven diverse languages, resulting in 805 unique tasks and 9,660 total language-specific instances.<n>We observe degradation in both performance and security when transitioning from English to other languages.
  arXiv  Detail & Related papers  (2025-05-21T18:42:00Z)
• Cross-Lingual Consistency: A Novel Inference Framework for Advancing Reasoning in Large Language Models [10.231866835957538]
  Chain-of-thought (CoT) has emerged as a critical mechanism for enhancing reasoning capabilities in large language models (LLMs)<n>We propose the Cross-Lingual Consistency (CLC) framework, which integrates multilingual reasoning paths through majority voting to elevate LLMs' reasoning capabilities.<n> Empirical evaluations on the CMATH dataset reveal CLC's superiority over the conventional self-consistency method.
  arXiv  Detail & Related papers  (2025-04-02T16:09:39Z)
• How Does Quantization Affect Multilingual LLMs? [50.867324914368524]
  Quantization techniques are widely used to improve inference speed and deployment of large language models. We conduct a thorough analysis of quantized multilingual LLMs, focusing on performance across languages and at varying scales.
  arXiv  Detail & Related papers  (2024-07-03T15:39:40Z)
• AM2iCo: Evaluating Word Meaning in Context across Low-ResourceLanguages with Adversarial Examples [51.048234591165155]
  We present AM2iCo, Adversarial and Multilingual Meaning in Context. It aims to faithfully assess the ability of state-of-the-art (SotA) representation models to understand the identity of word meaning in cross-lingual contexts. Results reveal that current SotA pretrained encoders substantially lag behind human performance.
  arXiv  Detail & Related papers  (2021-04-17T20:23:45Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.