arxiv arXiv cs.CL · just now · research

Qwen releases Qwen-CUA, a 397B-A17B native computer-use agent

from English

Researchers have introduced Qwen-CUA, a native computer-use agent built on a 397B-A17B Qwen mixture-of-experts backbone. The system operates by observing screenshots and acting through keyboard and mouse events, without relying on DOM trees or accessibility metadata.

It maintains up to 20 active screenshots and folds older visual history in fixed-size blocks to retain recent evidence.
Training utilized a cloud rollout fleet with nearly 100,000 vCPUs and approximately 40,000 verifiable tasks.
Qwen-CUA outperforms Qwen3.7 across eight benchmarks, reaching 86.2 on OSWorld-Verified and 18.5/48.4 binary/partial completion on OSWorld 2.0.
Scaling the approach to over one trillion parameters yields Qwen-CUA-Max, improving scores to 87.6 and 21.2/53.3 respectively.
The model reduces RedTeamCUA attack success from 36.6 to 16.4 relative to Qwen3.7.

These results establish native computer use as a broadly capable agent foundation, highlighting scalable verifiable interaction and hybrid tool use as key directions.

Importance 2/3 arXiv cs.CL Alibaba (Qwen) AI agents Research paper

Benchmarks

Benchmark	Model	Score
OSWorld 2.0	Qwen-CUA-Max	21.2%
OSWorld 2.0	Qwen-CUA	18.5%

Related articles

arxiv arXiv cs.AI · 4d ago · 1 view

Qwen-UI-Agent sets state of the art on mobile-use benchmarks

The Qwen team has released a technical report for Qwen-UI-Agent, a real-world centric foundation GUI agent designed to operate reliably across mobile, computer-use, web, and DeepSearch environments. The system combines diverse sandbox environments with a large-scale real-device mobile runtime, interleaving GUI operations with CLI execution and supporting long-horizon tasks.

arxiv arXiv cs.CL · 7d ago

AgentOmnia scales agentic models across full-scenario applications

Researchers present AgentOmnia, a framework for full-scenario agentic scaling that coordinates task-space definition, data synthesis, post-training, evaluation, and improvement across To-Consumer, To-Business, and To-Employee applications. The system utilizes an extensible taxonomy and constructs 5,018 stateful environments with 255,375 tools and 52,361 tasks to enable fine-grained diagnosis via OmniaBench.

arxiv arXiv cs.LG · 19d ago

TRACE improves long-horizon agent tool-use via turn-level credit estimation

The authors propose TRACE (Turn-level Reward Assignment via Credit Estimation), a dense credit-assignment method for agentic reinforcement learning that addresses the sparsity and high variance of outcome rewards in long-horizon tasks. By representing rollouts as state transitions at tool-call boundaries and deriving per-action rewards from log-ratio state values, TRACE enables effective training without additional critics or supervised fine-tuning stages.

arxiv arXiv cs.AI · 40d ago · 1 view

Hypothesis-Driven Skill Optimization for LLM Agents

HDSO enables safe, auditable skill updates for LLM agents without training, using falsifiable hypotheses and validation. On ALFWorld, it improves Qwen3-8B by +6.9 Avg. SR points and maintains a +7.1-point gain under noisy feedback, with validated skills transferable across runs and models when diagnostic alignment is achieved.

arxiv arXiv cs.CL · 41d ago · 6 views

Qwen-AgentWorld: Language World Models for General Agents

Qwen-AgentWorld-35B-A3B and Qwen-AgentWorld-397B-A17B are the first language world models that simulate agentic environments across seven domains using long chain-of-thought reasoning. Trained via a three-stage pipeline—CPT, SFT, and RL—these models outperform existing frontier models on AgentWorldBench, a benchmark derived from real-world interactions of five models on nine established tasks.