TextArena
MIT framework of 100+ text-based games for evaluating and training language-model agents, with a PyPI package and public site.
Tags
Description
Established open project (canonical repo TextArena/TextArena; not archived; created 2024-09-23; not presented as a 2026 launch). Root LICENSE is MIT (Copyright 2025 Leon Guertler and Bobby Cheng). PyPI textarena 0.7.4 verified 2026-11-13 (not yanked; wheels uploaded 2025-10-16; homepage https://www.textarena.ai/). The official install is pip install textarena. There were no GitHub Releases on 2026-11-13; the repo was last pushed 2026-10-05. The README on main describes later updates that post-date the PyPI upload, including 192-language support announced 2026-08-14; this entry does not claim those README updates are already in the 0.7.4 wheels. README and the public site describe 100+ single-, two-, and multi-player text games with a Gym-style interface for evaluation, self-play and RL. https://textarena.ai/ permanently redirects (308) to https://www.textarena.ai/ (HTTP 200); the www URL is used here. README examples use an OpenRouter API key; no keys were supplied and no games were run. Model-API and compute costs are separate.
Agent access
Related
- verifiers: Prime Intellect’s MIT library for creating LLM reinforcement-learning environments and evals, with a matching PyPI package.
- OpenEnv: BSD-3-Clause framework for building and serving isolated agentic RL environments over a Gymnasium-style reset/step/state API.
- NeMo Gym: NVIDIA Apache-2.0 library for building, evaluating and scaling agent environments, verifiers and harnesses, including sandboxed code-execution tasks.
Sources
- https://github.com/TextArena/TextArena
- https://github.com/TextArena/TextArena/blob/main/LICENSE
- https://github.com/TextArena/TextArena/blob/main/README.md
- https://pypi.org/project/TextArena/
- https://www.textarena.ai/
Machine-readable
- JSON
- Markdown
- Source file on GitHub (edit via pull request)