EWMBench

What is EWMBench?

EWMBench is a benchmark for evaluating Embodied World Models, covering aspects such as scene consistency, motion correctness, and semantic alignment. It…

Released

2025-05-16

Modality

multimodal

Evaluates

多模态, Multimodal, 智能体, Agent, 指令跟随, Instruct, worldmodel, embodied AI, manipulate, 物理智能, 指令遵循, Instruction Following, 具身交互, Embodied Interaction, 跨模态推理, Cross-modal Reasoning, 不支持, Unsupported

Openness

restricted

Importer

OpenCompass Hub

Reported scores

0

Source provenance

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

EWMBench paper, code and dataset

• Paper https://arxiv.org/abs/2505.09694
• Code https://github.com/AgibotTech/EWMBench
• Dataset https://huggingface.co/datasets/agibot-world/EWMBench

Which sources cite EWMBench?

• EWMBench

Related 多模态, Multimodal, 智能体, Agent, 指令跟随, Instruct, worldmodel, embodied AI, manipulate, 物理智能, 指令遵循, Instruction Following, 具身交互, Embodied Interaction, 跨模态推理, Cross-modal Reasoning, 不支持, Unsupported benchmarks

• 3DCodeBench
• A-Bench
• A-OKVQA
• AbsPyramid
• ActionAtlas
• ActiveFly-Bench
• Ada-LEval
• AECBench