Researchers have developed a post-training pipeline for large language models that enables them to achieve gold-medal performance in competitive programming. By combining supervised fine-tuning, reinforcement learning, and a new feedback-driven strategy called GenCorrect, the system outperforms top human contestants on the IOI 2026 problem set.

  • Nemotron-3-Nano-CC (30B-A3B) was trained with SFT and RL, while Nemotron-3-Ultra-CC (550B-A55B) used SFT alone.
  • The GenCorrect strategy iteratively generates, evaluates, and refines solutions to improve performance.
  • On IOI 2025, Nano-CC scored 468 with GenCorrect, exceeding the gold threshold of 438.3.
  • In IOI 2026, Ultra-CC scored 535.4 out of 600 under standard constraints, surpassing the top human score of 498.27.

This marks the first time an AI system has outscored the highest-scoring human contestant on an IOI problem set, demonstrating significant advances in reasoning and coding capabilities.