MINT-Bench: A Comprehensive Multilingual Benchmark for Instruction-Following Text-to-Speech

TL;DR

MINT-Bench is a new multilingual benchmark for instruction-following TTS that evaluates content, instruction adherence, and quality across ten languages, revealing current system limitations.

Contribution

It introduces a hierarchical, scalable benchmark and evaluation toolkit for multilingual, controllable TTS, addressing gaps in existing evaluation methods.

Findings

Current systems are far from fully solving instruction-following TTS.

Commercial systems outperform open-source models overall, but open-source models excel in Chinese.

Hard compositional and paralinguistic controls are major bottlenecks.

Abstract

Instruction-following text-to-speech (TTS) has emerged as an important capability for controllable and expressive speech generation, yet its evaluation remains underdeveloped due to limited benchmark coverage, weak diagnostic granularity, and insufficient multilingual support. We present \textbf{MINT-Bench}, a comprehensive multilingual benchmark for instruction-following TTS. MINT-Bench is built upon a hierarchical multi-axis taxonomy, a scalable multi-stage data construction pipeline, and a hierarchical hybrid evaluation protocol that jointly assesses content consistency, instruction following, and perceptual quality. Experiments across ten languages show that current systems remain far from solved: frontier commercial systems lead overall, while leading open-source models become highly competitive and can even outperform commercial counterparts in localized settings such as Chinese. The benchmark further reveals that harder compositional and paralinguistic controls remain major bottlenecks for current systems. We release MINT-Bench together with the data construction and evaluation toolkit to support future research on controllable, multilingual, and diagnostically grounded TTS evaluation. The leaderboard and demo are available at https://longwaytog0.github.io/MINT-Bench/