O
OSWorld
[NeurIPS 2024] OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments
Project.
Catalog snapshot
- License
- Apache 2.0
- Platforms
- Linux
- Docker
- No
- Commercial use
- Allowed
- Stars
- 3.1k
- Language
- Python
- Last commit
Repository health snapshot: 533 forks and 204 open issues at review time. Platform support is inferred conservatively from repository topics.
About.
[NeurIPS 2024] OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments Fossfits includes OSWorld as an actively maintained coding agents option: its public repository was updated on 2026-08-30, uses the APACHE-2.0 license, and had 3,140 stars when reviewed. Check its documentation and release notes before production adoption.
agent · artificial intelligence · benchmark · cli · code generation
