Benchmark

MMAR Benchmark

What is MMAR?

We introduce MMAR, a new benchmark comprising 1,000 meticulously curated audio-question-answer triplets, designed to evaluate the deep reasoning capabilities…

Released: 2025-05-11
Modality: multimodal
Evaluates: 强推理, Strong Reasoning, 多模态, Multimodal, 推理, Reasoning, MMAR, 音频语言模型, 深度推理, 多模态模型, VLM, 逻辑推理, 音频理解, Audio Understanding, 不支持, Unsupported
Openness: restricted
Importer: OpenCompass Hub
Reported scores: 0

Source provenance

Original evidence https://hub.opencompass.org.cn/dataset-detail/MMAR

MMAR paper, code and dataset

Paper https://arxiv.org/abs/2505.13032
Code https://github.com/ddlBoJack/MMAR
Dataset https://huggingface.co/datasets/BoJack/MMAR

No reported scores are on record for this benchmark yet.

Which sources cite MMAR?

MMAR

Related 强推理, Strong Reasoning, 多模态, Multimodal, 推理, Reasoning, MMAR, 音频语言模型, 深度推理, 多模态模型, VLM, 逻辑推理, 音频理解, Audio Understanding, 不支持, Unsupported benchmarks