NVIDIA's 550B coding model scores 535.4 on IOI 2026, first AI to beat top human contestant
jacek2023 · reddit · 2026-09-29
NVIDIA released Nemotron-Labs-3-Competitive-Coding (550B-A55B, NVFP4), a competitive-programming specialist fine-tuned from Nemotron-3-Ultra on 477,642 synthetic reasoning traces distilled from GLM-5.2 across 22,000 curated problems spanning 16 contest families. GLM-5.2 was chosen as SFT teacher for higher accuracy and 30% shorter generations than a DeepSeek-V4-Flash variant.
At inference, combined with GenCorrect, an iterative closed-loop test-time compute strategy that generates diverse candidates and refines them using evaluator feedback under a fixed submission budget, the model was evaluated prospectively on the IOI 2026 problem set under official time, internet, and submission constraints. It scored 535.4/600, beating both the gold-medal threshold (361.12) and the top human contestant (498.27) — the first reported AI system to outscore the highest-scoring human on an IOI set. Open weights, data, and recipes; commercially usable.
More from Models
- LLM shows surprisingly usable calibration classifying abstracts on human subjects — RexDouglass · 2026-09-29
- Dev flags suspected Opus 5.5 hardcoding of 'humans are right, AIs are wrong' bias — repligate · 2026-09-29
- Speculation: Meta paid full API prices for Fable traces to distill, and outputs taste like Claude — andersonbcdefg · 2026-09-29
- LastOPD: latent on-policy distillation collapses late, last-layer-only signal gains 5.55 on MATH-500 — Jie Yang · 2026-09-29
- Dev's take: OpenAI's $500 Pro plan is a bargain for client work, a hit for indie devs — alexcovo_eth · 2026-09-29
- Burkov questions whether Sonnet 5.5 matches Opus in Claude Code at half the cost — burkov · 2026-09-29