SoftwareGym synthesizes GUI environments for computer-use agents, lifting OSWorld 37% relative
_akhaliq · x · 2026-10-13
SoftwareGym is a new pipeline that automatically synthesizes diverse GUI software environments to train computer-use agents.
- Spans app categories like image editors, video players, spreadsheet editors, and code editors, with interaction chains of varying depth (opening images, brush strokes, importing CSVs, toggling dark mode)
- Delivers a 37% relative improvement on OSWorld, with Pass@3 rising from 41.4% to 57.0%
- The core idea: replace hand-built task suites with synthetic environments to scale training data for computer-use agents
Related event: SoftwareGym Auto-Synthesizes GUI Environments to Train Computer-Use Agents(2 posts)→
More from coding & agent
- PageIndex SDK v0.4 claims tree indexing now cheaper than vector indexing by ~15% — solyarisoftware · 2026-10-13
- Economist: agentic coding lets you write your own Fortran optimizers, not just ask for Stata packages — Apoorva__Lal · 2026-10-13
- Tencent open-sources Agent Memory: local long-term memory, 61% fewer tokens, 48%→76% accuracy — solyarisoftware · 2026-10-13
- Testing AI agents with simulations: replay real traces, 50 curated sims beat 1000 random ones — HamelHusain · 2026-10-13
- How humans and Copilot agents split the work to build a GitHub Copilot beginner course — DanWahlin · 2026-10-13
- Engineers distill their AGI debate into an essay on self-verifying, continuously improving agents — dotey · 2026-10-13