CMU's Gym-Anything Auto-Builds 200+ Agent Environments from Any Software, Selected by GDP
wellecks · x · 2026-10-30
CMU researchers released Gym-Anything, a framework that automatically converts arbitrary software into computer-use agent environments.
- Covers 200+ real applications (GIMP, Blender, OpenEMR, etc.), each installed, configured, and populated with realistic data in sandboxed VMs across Linux, Windows (via QEMU), and Android
- 50 tasks per software, generated via agentic seed creation and LLM amplification—10,000+ tasks total, each with checklist-based verifiers using privileged information
- Software selection is grounded in U.S. GDP decomposition, spanning all 22 SOC occupation groups
- Key insight: setting up agent environments is itself a coding/computer-use task, so the whole select-create-audit-scale pipeline can be automated
More from coding & agent
- PostTrainBench traces hit 104k monthly downloads — author warns it's likely reward hacking — maksym_andr · 2026-10-31
- Building games with AI? Set up a 'tech art' test scene to tune your look fast — chongdashu · 2026-10-31
- Skydive opens to all: watch your AI coworkers work in a 3D office — Scobleizer · 2026-10-31
- Handwritten RAGEN code becomes the #3 hardest task in FrontierCS 2 — wzenus · 2026-10-31
- LangChain SF meetup: decision models and a third path for agent evals — LangChain · 2026-10-31
- Cap your coding agent's retries at 5: how SuperPlane stops runaway token bills — Arindam_1729 · 2026-10-31