SoftwareGym: Scaling Computer-Use Agent Training with Diverse Synthesized Software Environments

Tian Jin; Wenhao Wang; Jing Kang; Haoting Shi; Rui Ye; Yijun Lu; Siheng Chen

cs.MAOct 8, 2026

SoftwareGym: Scaling Computer-Use Agent Training with Diverse Synthesized Software Environments

Authors: Tian Jin, Wenhao Wang, Jing Kang, Haoting Shi, Rui Ye, Yijun Lu, Siheng Chen

Organizations: Shanghai Jiao Tong University · Zhejiang University

Abstract

Computer-use agents (CUAs) are rapidly improving, yet synthesizing diverse and reliable software environments at scale remains challenging. Native application environments require substantial construction and integration effort, which limits application coverage and increases the risk of application-specific overfitting. Browser-hosted environments provide limited interaction depth and workflow complexity. These limitations motivate SoftwareGym, an automated pipeline for synthesizing GUI software environments with diverse interfaces, executable functionality, and runtime-verified behavior. It generates structurally diverse designs, realizes them as runnable applications, and uses GUI evaluation with iterative repair to ensure functional reliability. Using SoftwareGym, we synthesize 182 applications across twelve categories and achieve a mean GUI task pass rate of 91.5%. We construct SoftwareGym-Bench, a 120-task evaluation suite spanning the same twelve categories, and use trajectories collected from these environments for CUA training. The resulting model improves Pass@3 from 41.39% to 56.97% on the strictly out-of-domain OSWorld benchmark and achieves 40.0% on SoftwareGym-Bench. These results demonstrate the potential of synthesized software environments as scalable and transferable supervision for CUA training.

Figures & tables

Figure 1: Representative applications synthesized by SoftwareGym. Pic-Editor, VideoPlayer, TableEditor, and CodeEditor represent diverse software categories and GUI designs. Each row traces a representative multi-step workflow, demonstrating substantial interaction depth.

Figure 2: SoftwareGym synthesizes diverse, reliable, and executable applications through design, implementation, and evaluation. Design produces rich functional specifications and distinct interfaces. Implementation realizes and verifies the applications. Evaluation tests their GUI behavior and returns failures for subsequent implementation repair.

Figure 3: SoftwareGym grounds every synthesized task in capabilities verified through GUI interaction. It first extracts evidence-supported capabilities, then assigns coherent combinations to difficulty- and state-controlled slots, and finally realizes each slot as a natural request with an explicit initial state and expected outcome.

Panel of Table 1 — Table 1: Pass@3 (%) on OSWorld. Within each model category, bold and underlined scores denote the best and second-best results, respectively.

Panel of Table 2 — Table 2: Pass@3 (%) on SoftwareGym-Bench. Within each model category, bold and underlined scores denote the best and second-best results, respectively.

Panel of Figure 6 — Figure 6: Iterative evaluation and implementation repair consistently improve software reliability across the synthesized application library.

Figure 7: Structured style design improves blueprint diversity, while the GUI task pass rate under GUI-guided repair rises across both rounds and remains near its start under source-only repair.

Appendix figures & tables10 assets

Supplementary material from the paper’s appendix.

Appendix

Panel of Figure 8 — Figure 8: A successful TableEditor interaction that replaces “Draft total” in Sheet1!B2 with “Final total” while leaving the Find and Replace panel open.

Panel of Figure 9 — Figure 9: A successful PicEditor interaction that processes two images as JPEG files at 640-pixel width and quality 72, leaves both queue rows marked Done, and keeps the Batch Queue page open.

Panel of Figure 10 — Figure 10: A successful MusicPlayer interaction that enables the favorites-only library filter and leaves the filtered Library view open.

Panel of Figure 11 — Figure 11: A successful CodeEditor interaction that removes the leftover rainbow-indent extension while preserving the enabled built-in extensions.

Panel of Figure 12 — Figure 12: A successful ForumAPP interaction that creates the public, open-join Civic Desk community with the requested summary.

Panel of Figure 13 — Figure 13: A successful ChatAPP interaction that attaches project_brief.txt to the Loom Council conversation.

Category	Full	Simple	Category	Full	Simple
Browser	87.5	1.8	Mail	83.9	3.8
Chat	81.3	1.6	Rich-text Editor	85.3	2.4
Code Editor	85.8	3.6	Slides	84.4	5.7
Forum	82.1	5.4	Image Editor	81.5	21.1
Spreadsheet	78.4	2.6	Video Player	86.6	4.3
Music Player	81.5	3.1	PDF	85.5	3.4

Appendix

Table 4: Full Design consistently produces more diverse blueprint structures across all twelve categories. Results are on a 0–100 scale; each category contains one set of ten blueprints per method.

Method	$v_{1}$	$v_{2}$	$v_{3}$
Evaluation-guided	50.4 [42.6, 57.9]	70.5 [64.7, 76.2]	83.1 [79.2, 86.8]
Source-only	50.4 [42.6, 57.9]	50.5 [43.1, 57.7]	48.8 [42.7, 54.7]

Appendix

Table 5: Evaluation-guided implementation repair substantially improves functional reliability, whereas source-only repair remains near the shared starting level. Macro-average GUI task pass rate with 95% hierarchical bootstrap confidence intervals.

Model	Run	Chrome	GIMP	Calc	Impress	Writer	Thunderbird	VLC	VS Code	Overall
Seed 2.1 Pro	Run 1	67.26	46.15	87.23	73.41	69.56	86.67	80.49	69.57	72.59
	Run 2	58.56	50.00	76.60	71.28	82.60	86.67	86.37	65.22	70.13
	Run 3	65.08	61.54	70.21	64.90	73.90	86.67	74.61	65.22	68.49
	Avg@3	63.63	52.56	78.01	69.86	75.35	86.67	80.49	66.67	70.41
	Pass@3	71.74	73.08	93.62	74.47	78.26	86.67	76.47	82.61	79.51
Seed 2.0 Pro	Run 1	60.73	57.69	68.09	64.90	69.16	93.33	80.49	69.57	67.64

Appendix

Table 8: Run 1–3 scores, Avg@3, and Pass@3 (%) on OSWorld.

Model	Run	Pic	Mail	PDF	Music	Code	Forum	Writer	Browser	Video	Slides	Table	Chat	Overall
GPT-5.4	Run 1	60.00	80.00	80.00	40.00	80.00	60.00	30.00	70.00	70.00	50.00	60.00	70.00	62.50
	Run 2	80.00	80.00	70.00	40.00	70.00	60.00	50.00	90.00	60.00	50.00	40.00	80.00	64.17
	Run 3	70.00	90.00	90.00	50.00	70.00	60.00	30.00	80.00	50.00	50.00	60.00	70.00	64.17
	Avg@3	70.00	83.33	80.00	43.33	73.33	60.00	36.67	80.00	60.00	50.00	53.33	73.33	63.61
	Pass@3	80.00	90.00	90.00	50.00	80.00	80.00	50.00	90.00	70.00	60.00	70.00	80.00	74.17
GPT-5.5	Run 1	70.00	90.00	90.00	50.00	60.00	70.00	40.00	80.00	50.00	40.00	70.00	70.00	65.00

Appendix

Table 9: Run 1–3 scores, Avg@3, and Pass@3 (%) on SoftwareGym-Bench.

Computer-Use Agents

Computer-use agents perceive and operate graphical interfaces and software tools to complete tasks across websites, apps, and desktops.

See all 155 papers

Gym-Anything: Turn any Software into an Agent Environment
Apr 7, 2026Pranjal Aggarwal, Graham Neubig, Sean WelleckComputer-Use Agents Computer-Use Agent Benchmarks
CMU
CUA-Gym: Scaling Verifiable Training Environments and Tasks for Computer-Use Agents
May 25, 2026Bowen Wang, Dunjie Lu, Junli Wang +11Computer-Use Agents Synthetic Data Generation
The University of Hong Kong · Qwen Team, Alibaba Inc. · University of California, San Diego +1
Agent World Model in the Wild: Computer-Use Agent Environment Synthesis
Oct 19, 2026Zhaoyang Wang, Qianhui Wu, Haorui Wang +9Computer-Use Agents Synthetic Data Generation for LLM Agents
UNC-Chapel Hill · Microsoft Research · Georgia Tech +2

Apr 7, 2026J/K move · Enter open · S save

Gym-Anything: Turn any Software into an Agent Environment

Pranjal Aggarwal, Graham Neubig, Sean Welleck

CMU

Computer-Use Agents Computer-Use Agent Benchmarks Computer-Use Agent Evaluation Long-Horizon Agent Evaluation

Open paper AskPDF arXiv

Computer-use agents hold the promise of assisting in a wide range of digital economic activities. However, current research has largely focused on short-horizon tasks over a limited set of software with limited economic value, such as basic e-commerce and OS-configuration tasks. A key reason is that creating environments for complex software requires significant time and human effort, and therefore does not scale. To address this, we introduce Gym-Anything, a framework for converting any software into an interactive computer-use environment. We frame environment creation itself as a multi-agent task: a coding agent writes setup scripts, downloads real-world data, and configures the software, while producing evidence of correct setup. An independent audit agent then verifies evidence for the environment setup against a quality checklist. Using a taxonomy of economically valuable occupations grounded in U.S. GDP data, we apply this pipeline to 200 software applications with broad occupational coverage. The result is CUA-World, a collection of over 10K long-horizon tasks spanning domains from medical science and astronomy to engineering and enterprise systems, each configured with realistic data along with train and test splits. CUA-World also includes CUA-World-Long, a challenging long-horizon benchmark with tasks often requiring over 500 steps, far exceeding existing benchmarks. Distilling successful trajectories from the training split into a 2B vision-language model outperforms models 2

\times

its size. We also apply the same auditing principle at test time: a separate VLM reviews completed trajectories and provides feedback on what remains, improving Gemini-3-Flash on CUA-World-Long from 11.5% to 14.0%. We release all code, infrastructure, and benchmark data to facilitate future research in realistic computer-use agents.

Abstract

Computer-Use Agents

Explore similar work

Gym-Anything: Turn any Software into an Agent Environment

CUA-Gym: Scaling Verifiable Training Environments and Tasks for Computer-Use Agents

Agent World Model in the Wild: Computer-Use Agent Environment Synthesis