Dysca: A Dynamic and Scalable Benchmark for Evaluating Perception Ability of LVLMs

Jie Zhang, Zhong Ling Wang, Mengqi Lei, Zheng Yuan, Bei Yan +2 more
2/12/2026

No official code found. Best implementation: Benchmark-Dysca/Dysca (6 stars, last commit 2025-01-24). 1 more below.

Semantic Scholar

Code Implementations(2)

Robin-WZQ/DyscaCommunity73%

[ICLR'25] Dysca: A Dynamic and Scalable Benchmark for Evaluating Perception Ability of LVLMs

50PythonDec 9, 20241 years agoNOASSERTION
benchmarklvlmsperception-ability
Benchmark-Dysca/DyscaCommunity67%

Dysca: A Dynamic and Scalable Benchmark for Evaluating Perception Ability of LVLMs

60PythonApr 17, 20241 years agoNOASSERTION
benchmarklvlms

Cite this paper

@article{zhang2026dysca,
  title  = {Dysca: A Dynamic and Scalable Benchmark for Evaluating Perception Ability of LVLMs},
  author = {Jie Zhang and Zhong Ling Wang and Mengqi Lei and Zheng Yuan and Bei Yan and Shiguang Shan and Xilin Chen},
  year   = {2026},
  url    = {https://api.semanticscholar.org/CorpusID:7f7e2fc4a7920256d80a539d4f4ef3437c1dbe22},
  journal = {ICLR 2025 2025}
}

Discussion