Nemotron Fine-Tuned Models Achieve Gold-Medal-Level Results at IOI and IMO
Understand the story
On October 7, 2026, the Hugging Face Blog published a first-hand report that NVIDIA's team fine-tuned models based on Nemotron 3 through SFT, reinforcement learning, and a generate-verify-improve reasoning system, achieving gold-medal level in two competitions: scoring 30/42 at IMO 2026 (gold-medal threshold: 29 points); and in unofficial testing at IOI 2026, Nemotron-3-Ultra-CC scored 535.4/600 (gold-medal threshold: 361.12 points). The team also open-sourced the related recipes. There is only this one report on the event so far, with no follow-up developments.
AI overview of reporting · Updated 11 days ago
Reporting timeline
Explore different aspects of the story through its reporting.
- Hugging Face BlogSelectedOne Model Family, Two Gold-Level Results: Fine-Tuning Nemotron for IOI and IMO
The NVIDIA team, building on Nemotron 3, used SFT, RL, and a generate-verify-improve reasoning system to score 30/42 at IMO 2026 (gold threshold 29), while Nemotron-3-Ultra-CC scored 535.4/600 in unofficial IOI 2026 testing (gold threshold 361.12).