@inproceedings{yongsatianchot-etal-2026-fahmai,
title = "{F}ah{M}ai-{T}ele{B}ench: A {T}hai Grounded Tool-Use Benchmark for the Deployable Model Tier",
author = "Yongsatianchot, Nutchanon and
Ittichaiwong, Piyalitt and
Sawatphol, Jitkapat and
Ruangsiri, Tichanat",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the {A}sia-Pacific Chapter of the {A}ssociation for {C}omputational {L}inguistics (Volume 2: Short Papers)",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.aacl-short.15/",
pages = "185--196",
ISBN = "979-8-89176-440-8",
abstract = "Multilingual agentic benchmarks remain rare, especially for under-resourced languages such as Thai, where nicknames, unspaced script, and fused honorifics change how an agent must query its tools. We present FahMai-TeleBench, a bilingual (Thai/English) benchmark for grounded tool use. A model answers employee-directory questions over a synthetic company directory by issuing tool calls, or refuses when the answer is absent. All 626 items are synthetic but modeled on a real Thai enterprise-directory deployment, and they are graded deterministically without an LLM judge. Our central design choice is to treat tool availability as an experimental variable: each item runs under four tool configurations (grep, structured search, both, and a Python REPL), so the benchmark measures not just whether a model answers but how it uses tools. The task is near-solved at the frontier (GPT-5.4 97{--}99{\%}), yet sharply discriminating for the Thai-capable models an organization would host itself (Typhoon-2.5 58{--}74{\%}, Typhoon-S-8B 25{--}37{\%}). The Thai penalty is model-specific, and Typhoon-2.5 answers 90{\%} of items correctly once the relevant records are placed in its context, so its failures are failures of tool use. The tool axis also exposes differences at the frontier: GPT-5.5 drops to 73{\%} with a wide-schema structured search tool while scoring 97{--}98{\%} with grep and the REPL."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yongsatianchot-etal-2026-fahmai">
<titleInfo>
<title>FahMai-TeleBench: A Thai Grounded Tool-Use Benchmark for the Deployable Model Tier</title>
</titleInfo>
<name type="personal">
<namePart type="given">Nutchanon</namePart>
<namePart type="family">Yongsatianchot</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Piyalitt</namePart>
<namePart type="family">Ittichaiwong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jitkapat</namePart>
<namePart type="family">Sawatphol</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tichanat</namePart>
<namePart type="family">Ruangsiri</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 2: Short Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-440-8</identifier>
</relatedItem>
<abstract>Multilingual agentic benchmarks remain rare, especially for under-resourced languages such as Thai, where nicknames, unspaced script, and fused honorifics change how an agent must query its tools. We present FahMai-TeleBench, a bilingual (Thai/English) benchmark for grounded tool use. A model answers employee-directory questions over a synthetic company directory by issuing tool calls, or refuses when the answer is absent. All 626 items are synthetic but modeled on a real Thai enterprise-directory deployment, and they are graded deterministically without an LLM judge. Our central design choice is to treat tool availability as an experimental variable: each item runs under four tool configurations (grep, structured search, both, and a Python REPL), so the benchmark measures not just whether a model answers but how it uses tools. The task is near-solved at the frontier (GPT-5.4 97–99%), yet sharply discriminating for the Thai-capable models an organization would host itself (Typhoon-2.5 58–74%, Typhoon-S-8B 25–37%). The Thai penalty is model-specific, and Typhoon-2.5 answers 90% of items correctly once the relevant records are placed in its context, so its failures are failures of tool use. The tool axis also exposes differences at the frontier: GPT-5.5 drops to 73% with a wide-schema structured search tool while scoring 97–98% with grep and the REPL.</abstract>
<identifier type="citekey">yongsatianchot-etal-2026-fahmai</identifier>
<location>
<url>https://aclanthology.org/2026.aacl-short.15/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>185</start>
<end>196</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T FahMai-TeleBench: A Thai Grounded Tool-Use Benchmark for the Deployable Model Tier
%A Yongsatianchot, Nutchanon
%A Ittichaiwong, Piyalitt
%A Sawatphol, Jitkapat
%A Ruangsiri, Tichanat
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 2: Short Papers)
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-440-8
%F yongsatianchot-etal-2026-fahmai
%X Multilingual agentic benchmarks remain rare, especially for under-resourced languages such as Thai, where nicknames, unspaced script, and fused honorifics change how an agent must query its tools. We present FahMai-TeleBench, a bilingual (Thai/English) benchmark for grounded tool use. A model answers employee-directory questions over a synthetic company directory by issuing tool calls, or refuses when the answer is absent. All 626 items are synthetic but modeled on a real Thai enterprise-directory deployment, and they are graded deterministically without an LLM judge. Our central design choice is to treat tool availability as an experimental variable: each item runs under four tool configurations (grep, structured search, both, and a Python REPL), so the benchmark measures not just whether a model answers but how it uses tools. The task is near-solved at the frontier (GPT-5.4 97–99%), yet sharply discriminating for the Thai-capable models an organization would host itself (Typhoon-2.5 58–74%, Typhoon-S-8B 25–37%). The Thai penalty is model-specific, and Typhoon-2.5 answers 90% of items correctly once the relevant records are placed in its context, so its failures are failures of tool use. The tool axis also exposes differences at the frontier: GPT-5.5 drops to 73% with a wide-schema structured search tool while scoring 97–98% with grep and the REPL.
%U https://aclanthology.org/2026.aacl-short.15/
%P 185-196
Markdown (Informal)
[FahMai-TeleBench: A Thai Grounded Tool-Use Benchmark for the Deployable Model Tier](https://aclanthology.org/2026.aacl-short.15/) (Yongsatianchot et al., AACL-IJCNLP 2026)
ACL
- Nutchanon Yongsatianchot, Piyalitt Ittichaiwong, Jitkapat Sawatphol, and Tichanat Ruangsiri. 2026. FahMai-TeleBench: A Thai Grounded Tool-Use Benchmark for the Deployable Model Tier. In Proceedings of the 15th International Joint Conference on Natural Language Processing and the 5th Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (Volume 2: Short Papers), pages 185–196, Hengqin, China. Association for Computational Linguistics.