Hugging Face's logo
TIGER-Lab 's Collections
WebsiteBench
StopLoss
SearchGen
FIM-Midtraining
ClawBench
RationalRewards
SWE-Next
SWE-QA-Pro
MMMU
VisPhyWorld
OpenResearcher
ImagenWorld
VisCoder2
BrowserAgent
EditReward
VideoScore2
Critique-Coder
WebExplorer
MMLU-Pro
VLM2Vec-V2
VisCoder
One-Shot-CFT
Pixel-Reasoner
MoCha
General-Reasoner
VL-Rethinker
Vamba
TheoremExplain
ABC
VisualWebInstruct
PixelWorld
AceCoder
CritiqueFineTuning
MAmmoTH-VL
ScholarCopilot
VISTA
OmniEdit
MEGA-Bench
VLM2Vec
TIGERScore
MAmmoTH
UniIR
ImagenHub
Science
StructLM
ConsistI2V
Mantis
MAmmoTH2
VideoScore
Long-Context

WebsiteBench

Can AI agents rebuild websites through browser exploration? The 70 WebsiteBench tasks for the Harbor framework.