ARC-AGI-1 (Semi-Private)
ARC-AGI-1 (Semi-Private) is run by ARC Prize Foundation. It has ranked 230 models, of which the catalogue holds 62, scoring from 0 to 98.5 on Score (%).
measured by ARC Prize Foundation · the board itself
| Place | Model | Metric | Score |
|---|---|---|---|
| 1stof 230 | Claude Fable 5 | Score (%) (max) | 98.5 |
| 3rdof 230 | GPT 6 Astra | Score (%) (xhigh) | 98.5 |
| 7thof 230 | Gemini 3.1 Pro Preview | Score (%) | 98 |
| 8thof 230 | GPT-5.6 Sol | Score (%) (xhigh) | 97.5 |
| 9thof 230 | Claude Opus 5 | Score (%) (high) | 97.5 |
| 11thof 230 | Claude Fable 5.1 | Score (%) (max) | 97.5 |
| 15thof 230 | GPT-5.5 Pro | Score (%) (high) | 96.5 |
| 17thof 230 | GPT-5.6 Terra | Score (%) (max) | 96.5 |
| 23rdof 230 | Gemini 3.7 Flash | Score (%) (high) | 95.5 |
| 27thof 230 | GPT-5.4 Pro | Score (%) (xhigh) | 94.5 |
| 29thof 230 | Kimi K3 | Score (%) (max) | 94.5 |
| 31stof 230 | Claude Opus 4.6 | Score (%) (high) | 94 |
| 33rdof 230 | GPT-5.4 | Score (%) (xhigh) | 93.67 |
| 34thof 230 | Claude Opus 4.7 | Score (%) (high) | 93.5 |
| 37thof 230 | Claude Opus 4.8 | Score (%) (max) | 92.5 |
| 38thof 230 | Gemini 3.5 Flash | Score (%) (high) | 92.5 |
| 47thof 230 | Gemini 3.6 Flash | Score (%) (high) | 91.17 |
| 51stof 230 | GPT-5.6 Luna | Score (%) (max) | 90.67 |
| 52ndof 230 | GPT-5.2 Pro | Score (%) (xhigh) | 90.5 |
| 54thof 230 | DeepSeek V4 Pro 0813 | Score (%) (low) | 90.5 |
| 58thof 230 | DeepSeek V4 Flash (0731) | Score (%) (max) | 89 |
| 62ndof 230 | Gemini 3 Deep Think | Score (%) | 87.5 |
| 63rdof 230 | Grok 4.6 | Score (%) (medium) | 87.5 |
| 64thof 230 | Grok 4.5 | Score (%) (medium) | 87.17 |
| 70thof 230 | Claude Sonnet 4.6 | Score (%) (high) | 86.5 |
| 81stof 230 | Inkling Small | Score (%) (xhigh) | 84 |
| 87thof 230 | Inkling | Score (%) | 79.5 |
| 95thof 230 | GLM 5.2 | Score (%) | 77 |
| 100thof 230 | o3 | Score (%) (low) | 75.7 |
| 101stof 230 | gemini-3-pro-preview | Score (%) | 75 |
| 104thof 230 | GPT 5.1 Thinking | Score (%) (high) | 72.83 |
| 107thof 230 | GPT-5 Pro | Score (%) | 70.17 |
| 111thof 230 | GPT-5 | Score (%) (high) | 65.67 |
| 113thof 230 | Kimi K2.5 | Score (%) | 65.33 |
| 115thof 230 | MiniMax M2.5 | Score (%) | 63.67 |
| 116thof 230 | GPT-5.4 mini | Score (%) (xhigh) | 63.67 |
| 119thof 230 | o3-pro | Score (%) (high) | 59.33 |
| 120thof 230 | o4-mini | Score (%) (high) | 58.67 |
| 126thof 230 | DeepSeek v3.2 | Score (%) | 57 |
| 131stof 230 | GPT-5 mini | Score (%) (high) | 54.33 |
| 134thof 230 | Gemini 3.5 Flash-Lite | Score (%) (high) | 53.5 |
| 136thof 230 | GPT-5.4 nano | Score (%) (xhigh) | 51.5 |
| 143rdof 230 | GLM-5 | Score (%) | 44.67 |
| 150thof 230 | Claude Sonnet 4 | Score (%) | 40 |
| 158thof 230 | Claude Opus 4 | Score (%) | 35.67 |
| 160thof 230 | o3-mini | Score (%) (high) | 34.5 |
| 163rdof 230 | Gemini 2.5 Flash | Score (%) | 33.33 |
| 166thof 230 | Gemini 2.5 Pro | Score (%) | 33 |
| 179thof 230 | Codex mini | Score (%) | 27.33 |
| 183rdof 230 | Claude Sonnet 4.5 | Score (%) | 25.5 |
| 190thof 230 | DeepSeek-R1 (0528) | Score (%) | 21.21 |
| 192ndof 230 | GPT-5 nano | Score (%) (medium) | 20.71 |
| 198thof 230 | Grok 3 Mini | Score (%) (low) | 16.5 |
| 203rdof 230 | Claude Haiku 4.5 | Score (%) | 14.33 |
| 204thof 230 | o1-mini | Score (%) | 14 |
| 211thof 230 | Qwen3-235B-A22B-Instruct-2507 | Score (%) | 11 |
| 218thof 230 | GPT-4.1 fine-tuned | Score (%) | 5.5 |
| 219thof 230 | Grok 3 | Score (%) | 5.5 |
| 222ndof 230 | Magistral Small 2506 | Score (%) | 5 |
| 223rdof 230 | GPT-4o (2024-11-20) | Score (%) | 4.5 |
| 226thof 230 | GPT-4.1 mini | Score (%) | 3.5 |
| 230thof 230 | GPT-4.1 nano | Score (%) | 0 |
Read from the board on 2026-09-16