KVoiceBench

KVoiceBench is a Korean spoken question answering benchmark for evaluating speech models, built upon VoiceBench.

Dataset Description

As there is no existing benchmark for evaluating Korean spoken question answering capabilities, we construct KVoiceBench by building upon the English benchmark VoiceBench.

Construction process:

1. All transcriptions from VoiceBench are translated into Korean.
2. Translations are normalized into speech-friendly text.
3. Audio is synthesized using a Qwen3-TTS system.
4. Linguistic features that do not exist in Korean (e.g., capitalization, certain grammatical rules) are removed or adapted to better align with the Korean language.

Subsets

The dataset consists of 6,928 samples across 9 subsets:

Subset	Samples	Task Type
OpenBookQA	445	Multiple-choice QA
MMSU	3,006	Multiple-choice QA
Big-Bench-Hard	750	Multiple-choice QA
SD-QA	533	Short-answer QA
AlpacaEval	614	Open-ended
CommonEval	194	Open-ended
WildVoice	972	Open-ended
IFEval	279	Instruction following
AdvBench	513	Safety

Features

• id (string): Unique sample identifier
• audio (Audio): Synthesized Korean speech audio
• transcription (string): Korean text transcription of the audio
• answer (string): Reference answer (empty string for open-ended/safety subsets)
• capability (string): Evaluated capability category

Usage

from datasets import load_dataset

# Load a specific subset
ds = load_dataset("KRAFTON/KVoiceBench", "bbh-test", split="test")
print(ds[0])
# {'id': 'bbh_navigate_147', 'audio': {'path': ..., 'array': array([...]), 'sampling_rate': 24000},
#  'transcription': '...', 'answer': '...', 'capability': 'context_related/multi_hop_reasoning'}

# Load all subsets
subsets = [
    "advbench-test", "alpacaeval_full-test", "bbh-test", "commoneval-test",
    "ifeval-test", "mmsu", "openbookqa-test", "sd-qa", "wildvoice-test"
]
for subset in subsets:
    ds = load_dataset("KRAFTON/KVoiceBench", subset, split="test")
    print(f"{subset}: {len(ds)} samples")

Citation

@article{kvoicebench2026,
  title={KVoiceBench: A Korean Spoken Question Answering Benchmark},
  author={},
  year={2026}
}

License

This dataset is released under the Apache License 2.0.

© 2026 KRAFTON