RISEBench

What is RISEBench?

RISEBench is a benchmark for evaluating large multimodal models (LMMs) on reasoning-informed visual editing tasks, targeting models with image understanding…

Released

2025-04-03

Evaluates

推理, Reasoning, 指令跟随, Instruct, 多模态模型, VLM, 视觉生成, Visual Generation, 指令遵循, Instruction Following, 不支持, Unsupported

Openness

restricted

Importer

OpenCompass Hub

Reported scores

0

Source provenance

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

RISEBench paper, code and dataset

• Paper https://arxiv.org/pdf/2504.02826
• Code https://github.com/PhoenixZ810/RISEBench
• Dataset https://huggingface.co/datasets/PhoenixZ/RISEBench

Which sources cite RISEBench?

• RISEBench

Related 推理, Reasoning, 指令跟随, Instruct, 多模态模型, VLM, 视觉生成, Visual Generation, 指令遵循, Instruction Following, 不支持, Unsupported benchmarks

• A-Bench
• A-OKVQA
• AbsPyramid
• ActionAtlas
• ActiveFly-Bench
• Ada-LEval
• AdvancedMathBench
• AECBench