Hugging Face BlogT1·Models
68 pts
NVIDIA fine-tuned the Nemotron model family to achieve gold-level performance in both the International Olympiad in Informatics (IOI) and the International Mathematical Olympiad (IMO). The results demonstrate that general-purpose LLMs, when domain-specifically tuned and composed with transparent inference workflows, can solve frontier human competition problems.