Share your thoughts, 1 month free Claude Pro on usSee more

Vision-Language In-Context Learning on VL-ICL Bench Fast Open MiniImageNet (test)

100Top-1 Accuracy

ReVisIT (GPT-4.1)

Updated 23d ago

Evaluation Results

Method	Links
ReVisIT (GPT-4.1) 2026.07		100
ReVisIT (GPT-4.1) 2026.07		100
ReVisIT (Doubao-1.5) 2026.07		100
ReVisIT (Doubao-1.5) 2026.07		99.5
LLaVA-OneVision-72B 2026.07		98.7
ReVisIT (Qwen3-VL-30B, open) 2026.07		98.5
ReVisIT (Doubao-1.5) 2026.07		97.5
ReVisIT (GPT-4.1) 2026.07		96.5
ReVisIT (Qwen3-VL-30B, open) 2026.07		95
ReVisIT (Qwen3-VL-30B, open) 2026.07		80
GPT-4V 2026.07		78
IDEFICS-80B 2026.07		62.5
Qwen-VL-Chat 2026.07		58
ReVisIT (Qwen3-VL-30B, open) 2026.07		19.5