MultiAgent-X: Multilingual Agentic Function-Calling Benchmark for 12 Languages

MultiAgent-X is a multilingual function-calling training and evaluation dataset targeting under-resourced languages. It contains 10,551 records across 12 languages, 7 writing systems, and 5 life-critical agentic domains, covering 1.3 billion speakers. The dataset was created by Saurabh-66 and last updated on May 3, 2026.

Use Cases

• Training multilingual function-calling models based on the described 10,551 records across 12 languages.
• Benchmarking agent performance in life-critical domains based on the five specified agentic domains.
• Evaluating AI model fairness and coverage for under-resourced languages based on the described coverage of 1.3 billion speakers.
• Researching cross-lingual transfer for agentic tasks based on the dataset's multilingual and multi-script structure.

Strengths

• Contains 10,551 records, providing a substantial base for training and evaluation.
• Covers 12 languages and 7 unique writing systems, offering linguistic diversity.
• Targets 5 life-critical agentic domains, which suggests practical application focus.
• Explicitly designed for under-resourced languages covering 1.3 billion speakers, addressing a noted gap.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known, but specific file formats and sample data are unavailable for inspection.
• Description metadata is limited; actual data quality and annotation consistency require manual inspection.

Provenance

Source

huggingface, created by Saurabh-66.

Collection Method

Created with Adaptive Data by Adaption; specific collection methodology is not detailed.

Freshness

Last updated 2026-05-03 15:28:02.