IOI — leaderboard

Based on the International Olympiad in Informatics.

Metric: Score (self-reported). Source: benchmarklist.com. Status: saturation imminent. 54 models tracked.

Top models

#ModelScore
1DeepSeek V4 Pro580.1
2Nemotron 3 Ultra570
3GLM-5.1456.5
4Qwen 3.5 397B A17B441.3
5Claude Fable 5 (Max)72.25
6GPT-5.4 (xHigh)67.83
7GPT-5.2 (xHigh)54.83
8Claude Opus 4.7 (Max)47.08
9Qwen 3.7 Max (Max)46.75
10GPT-5.3 Codex (xHigh)43.83
11Gemini 3 Flash (Preview) (High)39.08
12Gemini 3 Pro (High)38.83
13DeepSeek V4 Pro (Max)35.83
14Grok 4.2030.17
15Grok 426.17

Interactive version: theaggregate.ai/benchmark?slug=ioi · How the rankings work · Data refreshed daily, snapshot 2026-07-22.