Oriel — AI & technology worth noticing, every day.

Qwen/BabelArena

Dataset · https://huggingface.co/datasets/Qwen/BabelArena

BabelArena: A Large-Scale Multilingual Benchmark for LLM Agents BabelArena is a multilingual agent benchmark covering 23 languages and four agent benchmark families across 13 domains — 702 canonical tasks, 16,146 task instances. Every component an agent touches is localized: prompts, tasks, tool descriptions and the environments themselves. Our benchmark ensures component-complete adaptation, language diversity and verified… See the full description on the dataset page: https://huggingface.co/datasets/Qwen/BabelArena.

First seen by Oriel on .

AI summary

BabelArena is a 23-language benchmark for LLM agents covering 702 tasks across four families, with a released standard subset of 532 tasks and 12,236 instances.

  • The released data covers τ³-Bench, VitaBench, and DeepPlanning; ClawEval task data is distributed separately.
  • Each language localizes prompts, tasks, tools, simulators, databases, policies, retrieval corpora, and environments across 15 scripts and three resource levels.
  • BabelFlow translation achieved mean quality scores of 4.28, 4.02, and 3.91 for high-, medium-, and low-resource languages.
  • Evaluation reports Pass^1, Pass^3, and language consistency after three independent trials, and all five tested models performed worse in low-resource languages.
  • The frozen 5.2 GiB dataset is intended for evaluation, requires the BabelArena repository, and cannot run with --task-subset full.

#llm agents #multilingual evaluation #benchmarking #tool use

Generated by gpt-5.6-sol. Check the original for details.

Metric trends

Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained

Likes

Total likes reported by the source.

  • Hugging Face Datasets: 9, latest observation ; 0 over observed period since

Downloads (30d)

Rolling 30-day downloads. A change in this value is not new downloads during the observed period.

  • Hugging Face Datasets: 9,158, latest observation ; +32 over observed period since

Aliases

SourceExternal IDURL
Hugging Face DatasetsQwen/BabelArenahttps://huggingface.co/datasets/Qwen/BabelArena Read discussion