IOI 2024 (Vals v2)
Code
The six IOI 2024 problems under Vals' v2 OpenCode protocol without interactive grading feedback. Final solutions receive official subtask points, averaged across the problems.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5 | 100.0 |
| Gemini 4 Argon | 100.0 |
| GPT-5.6 Sol | 100.0 |
| GPT-6.1 Sol | 100.0 |
| GPT-6 Astra | 100.0 |
| GPT-6 Sol | 100.0 |
| Kimi K3.1 Pro | 100.0 |
| Claude Sonnet 5.5 | 85.0 |
| GPT-5.6 Luna | 85.0 |
| GPT-5.6 Terra | 85.0 |
| Qwen3.8-Max | 78.2 |
| GLM-5.3 | 76.3 |
| Muse Wonder Full | 75.3 |
| Gemini 3.7 Flash | 72.8 |
| GPT-6 Luna | 69.3 |
| Claude Haiku 5.5 | 67.7 |
| Muse Wonder | 66.8 |
| Ember-1 | 64.5 |
| GPT-5.3-Codex | 62.8 |
| Grok 4.7 | 61.7 |
| Kimi K3 | 59.7 |
| Muse Spark 1.3 | 56.0 |
| Gemini 3.8 Flash | 55.7 |
| Gemini 3.1 Pro Preview | 55.5 |
| Hy4 Preview | 54.2 |
| GLM-5.3 Flash | 53.2 |
| Sonnet 5 | 50.5 |
| MiMo-V2.6-Flash | 50.5 |
| Grok 4.6 | 50.3 |
| Mistral Large 4 Preview | 50.2 |
| DeepSeek V4 Pro 0813 | 50.0 |
| Grok 4.5 | 47.2 |
| MiMo-V2.6-Pro | 43.5 |
| DeepSeek V4.1 Flash | 42.7 |
| Gemini 3.6 Flash | 40.7 |
| DeepSeek-V4-Flash-0731 | 34.0 |
| Qwen3.8 27B | 31.8 |
| Step 5 Preview | 26.7 |
| Muse Spark 1.2 | 22.0 |
| Inkling-Small | 17.2 |