IOI - IOI 2025
Code · 2025-08-11
The IOI 2025 split of IOI. This child benchmark separates a source-reported subtask or subtrack from the parent aggregate so scores at different grains do not share one benchmark_slug.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-5.6 Sol | 88.3 |
| Claude Opus 5 | 83.3 |
| GPT-5.4 | 68.5 |
| GPT-5.6 Luna | 67.8 |
| GPT-5.2 | 65.8 |
| GPT-5.6 Terra | 64.3 |
| Qwen3.7-Max | 53.0 |
| Claude Fable 5 | 44.5 |
| Opus 4.7 | 41.0 |
| DeepSeek-V4-Pro | 40.3 |
| Gemini 3 Pro Preview | 39.2 |
| Grok 4.20 | 37.8 |
| Gemini 3 Flash Preview | 37.2 |
| GPT-5.1 | 31.2 |
| GPT-5.3-Codex | 29.5 |
| Grok 4 | 29.2 |
| GPT-5 | 29.0 |
| GPT-5.1-Codex-Max | 28.3 |
| Gemini 3.5 Flash-Lite | 27.3 |
| Opus 4.5 | 26.5 |
| Kimi K2.5 | 23.2 |
| GLM-5 | 20.3 |
| Sonnet 4.5 | 17.7 |
| Grok 4.3 | 17.7 |
| GPT-5-Codex | 16.7 |
| GPT-5.4 Nano | 16.3 |
| Gemini 2.5 Pro | 15.2 |
| Qwen3 Max Thinking (2026-01-23) | 15.2 |
| DeepSeek-V3.2 | 14.8 |
| Grok 4 Fast | 13.7 |
| Qwen3 Max (rolling alias) | 9.8 |
| MiniMax M2.7 | 9.3 |
| GLM 4.6 | 7.7 |
| Grok 4.1 Fast | 7.7 |
| GPT-5.1-Codex | 7.3 |
| GPT-5 Mini | 7.0 |
| Opus 4.1 | 6.4 |
| GPT-5.4 Mini | 6.2 |
| Haiku 4.5 | 5.8 |
| MiniMax M2.5 | 5.8 |