Benchmark

GSM8K-V Benchmark

What is GSM8K-V?

GSM8K-V is a multi-image, purely visual mathematical reasoning benchmark built by rendering GSM8K textual problems into visual scenes. It exposes substantial…

Released: 2025-09-28
Modality: multimodal
Evaluates: 多模态, Multimodal, 推理, Reasoning, 数学, Math, visual mathematical reasoning, gsm8k, multi-image math benchmark, 多模态模型, VLM, 逻辑推理, 数理能力, 不支持, Unsupported
Openness: open
Importer: OpenCompass Hub
Reported scores: 0

Source provenance

Original evidence https://hub.opencompass.org.cn/dataset-detail/GSM8K-V

GSM8K-V paper, code and dataset

Paper https://arxiv.org/abs/2509.25160
Code https://github.com/ZJU-REAL/GSM8K-V
Dataset https://huggingface.co/datasets/ZJU-REAL/GSM8K-V

No reported scores are on record for this benchmark yet.

Which sources cite GSM8K-V?

GSM8K-V

Related 多模态, Multimodal, 推理, Reasoning, 数学, Math, visual mathematical reasoning, gsm8k, multi-image math benchmark, 多模态模型, VLM, 逻辑推理, 数理能力, 不支持, Unsupported benchmarks