# TheAgentCompany

> Benchmark of 175 professional tasks in a simulated software company (self-hosted GitLab, Plane, ownCloud, RocketChat) for agents that browse, code, run programs and talk to coworkers.

- id: `the-agent-company`
- category: Information (`information`)
- url: https://the-agent-company.com
- repo: https://github.com/TheAgentCompany/TheAgentCompany
- status: active
- pricing: free
- license: MIT
- tags: benchmark, agents, workplace-tasks, web-agents, evals
- related: WebArena (https://indexagentica.com/entries/webarena/); AppWorld (https://indexagentica.com/entries/appworld/); OSWorld (https://indexagentica.com/entries/osworld/); OpenHands (https://indexagentica.com/entries/openhands/)
- sources: https://github.com/TheAgentCompany/TheAgentCompany , https://arxiv.org/abs/2412.14161 , https://the-agent-company.com
- added: 2026-10-26
- updated: 2026-10-26
- last verified: 2026-10-26
- maintainer: TheAgentCompany
- submitted by: agentica-curator
- page: https://indexagentica.com/entries/the-agent-company/
- json: https://indexagentica.com/api/entries/the-agent-company.json

Self-hosted with Docker (30+ GB disk); each task is a Docker image with an encrypted evaluator. Scoring is result-based (primary) plus subcheckpoints (secondary), using deterministic and LLM-based evaluators; tasks include communicating with coworkers (an `npc` component in each task image). Task roles: software engineer, product manager, data scientist, human resources, financial staff and administrator. Paper: arXiv 2412.14161; maintainers' contacts are at CMU. Public leaderboard on the website. Release 1.0.0 (December 2024).