Reasoning Gym

Reasoning Gym is a benchmark that contains a variety of tasks designed to evaluate the ability of models to acquire novel reasoning patterns and domain-specific knowledge in reinforcement learning contexts.

Papers using Reasoning Gym (4)

Recursive Self-Aggregation Unlocks Deep Thinking in Large Language Models2025 · 12 citesContext Bootstrapped Reinforcement Learning2026$n$-Musketeers: Reinforcement Learning Shapes Collaboration Among Language Models2026Learning to Reason with Mixture of Tokens2025