Computer-Use Agents Function as Embodied Generalists
Computer-use agents win multiple ECCV competitions by treating embodied tasks as software interactions.
Stories per day · 30 days
2 RSS
Relatedqwen5osworld3agent-training2reinforcement-learning2synthetic-environments2action-replay1agent-evaluation1agent-interfaces1ai-agents1benchmarks1claude1cua-world1embodied-ai1game-agents1generalists1gui-agents1leash1llm-efficiency1
29 stories, newest first · one card per story
Stories tagged #computer-use-agents
arXiv introduces Visual Token Routing for efficient computer-use agents
VTR matches full-resolution performance on OSWorld and CUA-World using 2.5-5.8x fewer visual tokens, reducing inference cost by up to 3.3x.
CUA-Master scales computer-use agents via open-world recursive improvements
CUA-Master-27B scores 80.2 on OSWorld Verified and 36.9 on OSWorld 2.0 in GUI-only and hybrid benchmarks.
Agent S4 evolves interfaces and leads open-source OSWorld
With Qwen3.8-27B it scores 21.3% on OSWorld V2, 3.8 times the same model with fixed interfaces.
ReTrace framework studies recovery in computer-use agents
ReTrace reveals that current agents fail to exploit latent recovery capabilities despite goal-oriented action replay.
WildAWM synthesizes environments for computer-use agent training
The open-source pipeline made 341 long-horizon tasks; GPT-6-Astra solved 60%, and two trained Qwen models improved on eight benchmarks.
WildAWM synthesizes computer-use agent environments with open models
It yields 341 tasks that can exceed 100 steps; GPT-6-Astra solves 60%, and Qwen models trained only there improve on eight benchmarks.
PersonaCUABench: computer-use agents lose 36% completion on persona requests
Nine backbones ran 3,330 times; top satisfaction after full persona completion was 62.7% versus 84.1% on original requests.
OS-Helix trains computer-use agents to mix GUI and code
OS-Helix-27B-RL reaches 66.44% on OSWorld-Verified, a further 42.9% relative gain after RL.
LEASH: Uncertainty-Aware Weak-to-Strong Escalation and Oversight for Computer-Use Agents
LEASH escalates computer-use agents by latent uncertainty, matching Claude Opus 5 at 19% lower cost.
Player-1: Learning to Play from Synthetic Experience
Player-1 uses synthetic RL environments to substantially improve game-control agents beyond supervised training.
Work2Skill: Capturing Human GUI Experience for Reusable Agent Skill Induction
Work2Skill turns human GUI demos into reusable skills that raise computer-use agent scores.
SoftwareGym: Scaling Computer-Use Agent Training with Diverse Synthesized Software Environments
SoftwareGym's 182 synthetic apps lift a computer-use agent to 56.97% OSWorld Pass@3.
CUAWright: A Minimal Unified Interface for Digital Agents
CUAWright, a minimal bash-and-filesystem harness, outperforms GUI computer-use agents on several benchmarks.
DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents
DeskForge supplies 1.2 million annotated desktop scenes that improve GUI grounding and long-horizon computer-use agents.
cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents
cua-speedrun standardizes virtual-machine infrastructure to measure computer-use agent speed, cost, and efficiency.
ComputerSD: Online Self-Distillation from Real-Time Feedback for Computer-Use Agents
ComputerSD turns real-time GUI feedback into online self-distillation signals for computer-use agents.
OSWorld-Science: A Benchmark of Computer Use Agents for Learning and Using Scientific Software
OSWorld-Science benchmarks 12 VLMs on 146 scientific software tasks with artifact-based scoring.
CUA-SWE: When Computer-Use Agents Meet Visual Software Engineering
CUA-SWE tests agents that repair software using code changes and visual GUI feedback.
CaptchaArena: A Large-Scale, Fine-Grained Dataset for Training Computer-Use Agents on Interactive CAPTCHAs
CaptchaArena releases 50K verified interactive CAPTCHA trajectories and a 9B agent reaching 71.7 Pass@1.
OSWorld-Pro: Process-based Evaluation for Computer Use Agents
OSWorld-Pro adds process-level subgoals so computer-use agent failures can be diagnosed beyond final outcomes.
OSWorld-Pro: Process-based Evaluation for Computer Use Agents
OSWorld-Pro scores computer-use agents on over 300 tasks; Claude Opus 5 reaches 75.7%.
Show HN: CUA-S1 – A System One Model for Computer Use
Cua open-sources CUA-S1, a fast 'System One' model for driving computer-use agents, launched on GitHub via Show HN.
RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents
RecreationWorld benchmarks hybrid computer-use agents that rediscover and reimplement apps across five platforms.