Metraipolis › archive › dc_b978ced3e7a4

$τ$-Multilingual: Benchmarking Voice Agents Across Languages — arXiv:2609.35820

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.35820
License (as archived)
CC0-1.0
Archived
2026-10-07T16:24:47.529Z
SHA-256 of the archived text
25e3440c018603d07f37b512b4ac34d300136165d7bbe4513826f7cc57715734
Stable URL of this page
https://metraipolis.com/documento/dc_b978ced3e7a4

How to cite

$τ$-Multilingual: Benchmarking Voice Agents Across Languages — arXiv:2609.35820 — arXiv (en), https://arxiv.org/abs/2609.35820 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b978ced3e7a4 (Metraipolis archive dc_b978ced3e7a4, sha256 25e3440c018603d07f37b512b4ac34d300136165d7bbe4513826f7cc57715734)
BibTeX
@misc{dcb978ced3e7a4,
  title = {$τ$-Multilingual: Benchmarking Voice Agents Across Languages — arXiv:2609.35820},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.35820},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b978ced3e7a4; sha256 25e3440c018603d07f37b512b4ac34d300136165d7bbe4513826f7cc57715734},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35820. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

English-only benchmarks expose only a narrow slice of voice-agent behavior. We introduce $\tau$-Multilingual, extending $\tau$-Voice to Spanish, Brazilian Portuguese, Hindi, Korean, and Mandarin with native-speaker review and evaluation of generated language and spoken output. Across 4,500 full-duplex calls and five voice configurations, Spanish, Portuguese, and Hindi remain within 3.2 task-completion points of English, but Korean and Mandarin fall by 14.7 and 8.4 points. The failure modes also vary: Korean systems miss more responses, Mandarin systems interrupt more often, and both struggle with tools and entities. Grok leads task completion but scores lowest on generation quality, motivating separate task, interaction, and generation reporting. We release language packs, validated judges, and tools for community-built multilingual voice-agent evaluation.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Sound (cs.SD); Audio and Speech Processing (eess.AS).
Pagina del lavoro: https://arxiv.org/abs/2609.35820 — 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 2026-10-07T15:54:53.718Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).