Agent-S
Developed by simular-ai
Agent-S is an open-source framework by Simular AI designed for autonomous computer interaction via an Agent-Computer Interface (ACI). Its latest version, Agent S3, is the first to surpass human-level performance on the OSWorld benchmark with a score of 72.60%. The system operates by pairing high-level reasoning models with specialized grounding models like UI-TARS, enabling it to perform complex tasks across Windows, macOS, Linux, and Android. It features local code execution and the ability to learn from past experiences.
- Super-human OSWorld Performance
- Cross-OS GUI Interaction Support
- Agent-Computer Interface (ACI) Framework
- Local Python and Bash Execution
- Multi-model Grounding (e.g., UI-TARS)
desktopmobile