Qiguang Chen

Frontier LLMs can achieve substantial recursive self-improvement, closing nearly half of the performance gap on complex benchmarks by jointly optimizing data, training, and execution harnesses within a unified, agent-native environment.

Transforming scientific papers into multi-turn generation trajectories not only doubles the training data but also boosts academic writing benchmarks while maintaining reasoning skills.

LLM agents can enhance training-data strategies in over half of their attempts, but their inconsistency reveals critical limitations in recursive self-improvement.