Skip to main content
rLLM home page
Search...
⌘K
Ask AI
rllm-org/rllm
Slack
Slack
Search...
Navigation
Case studies
Cogito, Ergo Ludo
Documentation
Examples
Projects
API Reference
Blog
Case studies
FinQA Financial Agent
DeepScaleR
DeepCoder
DeepSWE
Tongyi DeepResearch
Terminal-Bench-RL
PettingLLMs
SETA
LLM-in-Sandbox
Cogito, Ergo Ludo
Cut the Bill, Keep the Turns
Experiential Reinforcement Learning
V1: Parallel Self-Verification
Case studies
Cogito, Ergo Ludo
An agent that learns to play games by reasoning and planning
Paper
arXiv:2509.25052
Cogito, Ergo Ludo combines reasoning and planning to build game-playing agents. Using rLLM’s RL training pipeline, it trains agents that can think through game states and develop strategies through self-play.
Was this page helpful?
Yes
No
LLM-in-Sandbox
Previous
Cut the Bill, Keep the Turns
Next
⌘I
Assistant
Responses are generated using AI and may contain mistakes.