OpenPipe ART
Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen3.6, GPT-OSS, Llama, and more!
OpenPipe ART scores 6.7/10 (Solid) in Fine-Tuning.
Language Python License Apache-2.0 Stars 10.2k★ Type library Self-host Yes Repo OpenPipe/ART Homepage https://art.openpipe.ai
Quality
Solid 6.7
Adoption 50
Activity 73
Maturity 85
Community 50
Capability 50
Show the math
Overall 6.7/10
value = 6.7/10
score = value
Adoption 5.5/10
value = 50.2/100
score = 1 + 9 * value / 100
Activity 7.5/10
value = 72.7/100
score = 1 + 9 * value / 100
Maturity 8.6/10
value = 84.8/100
score = 1 + 9 * value / 100
Community 5.5/10
value = 49.8/100
score = 1 + 9 * value / 100
Capability 5.5/10
value = 50/100
score = 1 + 9 * value / 100
Key metrics
10.2k★ GitHub stars
95/90d Recent commits
10 devs/90d Recent contributors
1.3y Project age
active Last commit
Gotchas
No gotchas documented yet. Contribute one if you know a constraint we missed.
Labels
- Self-hosted
Self-hosted
You deploy and operate it yourself; there is no hosted option here.
agent-rl grpo multi-turn agents reinforcement-learning