InfiGUIAgent: A Multimodal Generalist GUI Agent with Native Reasoning and Reflection

Open paper

Benchmark	Model	Rank	Results
natural-language-visual-grounding-on-screenspot	InfiGUIAgent-2B	#23	Accuracy (%): 76.3
computer-use-agents-on-androidworld	InfiGUIAgent-2B	#32	Success Rate: 9.0