CUAWright: A Minimal Unified Interface for Digital Agents
CUAWright, a minimal bash-and-filesystem harness, outperforms GUI computer-use agents on several benchmarks.
CUAWright is a roughly 3,000-line terminal harness that gives computer-use agents only bash commands and a filesystem for creating tools and managing context. On OSWorld 2.0 it reports a 33.2% relative improvement in partial reward and a 37.5% lower estimated cost versus the published GPT-5.5 baseline. It also beats GUI-native harnesses by 4.7 percentage points on Online-Mind2Web and 44.0 points on the Odysseys benchmark. On CADGenBench and BenchCAD, the harness yields 8.1% to 41.6% relative gains over other CLI harnesses with GPT-5.5.
- Sole action interface is bash; the filesystem stores evolving tools
- OSWorld 2.0: 33.2% relative partial-reward gain and 37.5% lower cost
- Beats GUI harnesses by 4.7% on Online-Mind2Web and 44.0% on Odysseys
- CAD benches improve 8.1% to 41.6% versus other CLI harnesses
Full article247 words · extracted from huggingface.co · click to collapse
The prevailing approach to computer-use agents couples a model with a domain-specific harness: a browser or desktop environment equipped with human engineered tools that are fixed before task execution. As models' coding capabilities improve, the GUI native and static harness prevents them from direct programmatic operation on system state, as well as flexible construction of tools. To this end, we introduce CUAWright, a minimal terminal harness of roughly 3K lines of code that uses bash commands as its sole action interface, and a file system as its evolvable space for dynamically creating tools and managing the context. We conduct comprehensive experiments across a wide range of digital tasks, and demonstrate that by giving the agent a minimal, programmable interface, it achieves substantially stronger results compared to their GUI or hybrid CLI interface across a wide range of tasks. On OSWorld 2.0, CUAWright delivers a 33.2% relative improvement in partial reward while reducing estimated cost by 37.5% compared with the published GPT-5.5 baseline. On Online-Mind2Web and the long horizon Odysseys benchmark, CUAWright substantially outperforms GUI native harness by 4.7% and 44.0% in success rate, respectively. Furthermore, we found the gains extend to CAD applications that require accurate visual understanding and CLI interaction: on CADGenBench and BenchCAD, our unified harness yields 8.1%-41.6% relative improvements over other CLI based harnesses with GPT-5.5. Together, these results suggest digital environments are far more programmable than their GUI interfaces imply, and a minimal terminal-focused harness is the key for better performance and efficiency.
Text extracted automatically; images, tables and formatting may be missing. Original: https://huggingface.co/papers/2610.04116