Researchers have introduced Qwen-CUA, a new AI agent designed for native computer interaction, capable of operating software using only screenshots and standard input devices. This agent utilizes a 397B-parameter Qwen mixture-of-experts model and maintains context by tracking up to 20 active screenshots. Developed on a large-scale cloud fleet with extensive task data, Qwen-CUA demonstrates strong performance across benchmarks, outperforming its predecessor Qwen3.7 and rivaling proprietary systems. A larger version, Qwen-CUA-Max, further improves these scores and significantly reduces vulnerability to red-teaming attacks. AI
IMPACT Establishes native computer use as a viable agent foundation, highlighting scalable interaction and hybrid tool use.
RANK_REASON Research paper detailing a new AI agent for computer interaction. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
