@inproceedings{yang-etal-2026-t2mbench,
title = "{T}2{MB}ench: Evaluating Text-to-Motion Generation under Diverse Language Instructions",
author = "Yang, Bin and
OU, Rong and
Xu, Weisheng and
Xiong, Jiaqi and
Li, Xintao and
Wang, Taowen and
Zhu, Luyu and
Fang, Yuetong and
Jiang, Xu and
Tan, Jing and
Xu, Renjing",
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 1: Long Papers)",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.aacl-long.75/",
pages = "1622--1662",
ISBN = "979-8-89176-439-2",
abstract = "As text-to-motion (T2M) models move beyond short action captions toward real-world applications, they are increasingly expected to follow diverse instructions. However, existing evaluation protocols remain insufficient for systematically assessing instruction following, physical plausibility, and fine-grained controllability under such diverse language instructions. To address this gap, we introduce T2MBench, a benchmark with 1,025 prompts across four categories{---}Dynamics, Complexity, Interaction, and Accuracy{---}covering 14 representative T2M models and three evaluation dimensions: LLM-Based Evaluation, Multi-Factor Motion Evaluation, and Fine-grained Accuracy Evaluation. Our results reveal substantial gaps in current T2M models' ability to consistently follow diverse language instructions while maintaining motion plausibility and fine-grained control.T2MBench establishes a reusable diagnostic benchmark for advancing research toward more reliable T2M generation under diverse language instructions. Data and materials are available at \url{https://github.com/Courteous121/T2MBench-release}."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yang-etal-2026-t2mbench">
<titleInfo>
<title>T2MBench: Evaluating Text-to-Motion Generation under Diverse Language Instructions</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bin</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Rong</namePart>
<namePart type="family">OU</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weisheng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiaqi</namePart>
<namePart type="family">Xiong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xintao</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Taowen</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Luyu</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuetong</namePart>
<namePart type="family">Fang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xu</namePart>
<namePart type="family">Jiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jing</namePart>
<namePart type="family">Tan</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Renjing</namePart>
<namePart type="family">Xu</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 1: Long 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">979-8-89176-439-2</identifier>
</relatedItem>
<abstract>As text-to-motion (T2M) models move beyond short action captions toward real-world applications, they are increasingly expected to follow diverse instructions. However, existing evaluation protocols remain insufficient for systematically assessing instruction following, physical plausibility, and fine-grained controllability under such diverse language instructions. To address this gap, we introduce T2MBench, a benchmark with 1,025 prompts across four categories—Dynamics, Complexity, Interaction, and Accuracy—covering 14 representative T2M models and three evaluation dimensions: LLM-Based Evaluation, Multi-Factor Motion Evaluation, and Fine-grained Accuracy Evaluation. Our results reveal substantial gaps in current T2M models’ ability to consistently follow diverse language instructions while maintaining motion plausibility and fine-grained control.T2MBench establishes a reusable diagnostic benchmark for advancing research toward more reliable T2M generation under diverse language instructions. Data and materials are available at https://github.com/Courteous121/T2MBench-release.</abstract>
<identifier type="citekey">yang-etal-2026-t2mbench</identifier>
<location>
<url>https://aclanthology.org/2026.aacl-long.75/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>1622</start>
<end>1662</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T T2MBench: Evaluating Text-to-Motion Generation under Diverse Language Instructions
%A Yang, Bin
%A OU, Rong
%A Xu, Weisheng
%A Xiong, Jiaqi
%A Li, Xintao
%A Wang, Taowen
%A Zhu, Luyu
%A Fang, Yuetong
%A Jiang, Xu
%A Tan, Jing
%A Xu, Renjing
%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 1: Long Papers)
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-439-2
%F yang-etal-2026-t2mbench
%X As text-to-motion (T2M) models move beyond short action captions toward real-world applications, they are increasingly expected to follow diverse instructions. However, existing evaluation protocols remain insufficient for systematically assessing instruction following, physical plausibility, and fine-grained controllability under such diverse language instructions. To address this gap, we introduce T2MBench, a benchmark with 1,025 prompts across four categories—Dynamics, Complexity, Interaction, and Accuracy—covering 14 representative T2M models and three evaluation dimensions: LLM-Based Evaluation, Multi-Factor Motion Evaluation, and Fine-grained Accuracy Evaluation. Our results reveal substantial gaps in current T2M models’ ability to consistently follow diverse language instructions while maintaining motion plausibility and fine-grained control.T2MBench establishes a reusable diagnostic benchmark for advancing research toward more reliable T2M generation under diverse language instructions. Data and materials are available at https://github.com/Courteous121/T2MBench-release.
%U https://aclanthology.org/2026.aacl-long.75/
%P 1622-1662
Markdown (Informal)
[T2MBench: Evaluating Text-to-Motion Generation under Diverse Language Instructions](https://aclanthology.org/2026.aacl-long.75/) (Yang et al., AACL-IJCNLP 2026)
ACL
- Bin Yang, Rong OU, Weisheng Xu, Jiaqi Xiong, Xintao Li, Taowen Wang, Luyu Zhu, Yuetong Fang, Xu Jiang, Jing Tan, and Renjing Xu. 2026. T2MBench: Evaluating Text-to-Motion Generation under Diverse Language Instructions. 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 1: Long Papers), pages 1622–1662, Hengqin, China. Association for Computational Linguistics.