ARC-AGI-2 (Semi-Private)
ARC-AGI-2 (Semi-Private) is run by ARC Prize Foundation. It has ranked 298 models, of which the catalogue holds 80, scoring from 0 to 95 on Score (%).
measured by ARC Prize Foundation · the board itself
| Place | Model | Metric | Score |
|---|---|---|---|
| 2ndof 298 | GPT 6 Astra | Score (%) (max) | 95 |
| 3rdof 298 | GPT 6.1 Sol | Score (%) (max) | 94.17 |
| 5thof 298 | Claude Opus 5.5 | Score (%) (high) | 93.33 |
| 6thof 298 | GPT-5.6 Sol | Score (%) (max) | 92.5 |
| 13thof 298 | Gemini 4 Argon | Score (%) (high) | 91.25 |
| 14thof 298 | Claude Opus 5 | Score (%) (max) | 90.42 |
| 16thof 298 | Claude Fable 5.1 | Score (%) (max) | 90 |
| 18thof 298 | GPT 6 Sol | Score (%) (max) | 89.58 |
| 19thof 298 | Claude Fable 5 | Score (%) (max) | 89.17 |
| 20thof 298 | Gemini 3.8 Flash | Score (%) (high) | 89.17 |
| 33rdof 298 | GPT-5.5 Pro | Score (%) (high) | 84.58 |
| 34thof 298 | Gemini 3.7 Flash | Score (%) (high) | 84.58 |
| 36thof 298 | GPT-5.6 Terra | Score (%) (max) | 83.9 |
| 37thof 298 | Claude Sonnet 5.5 | Score (%) (xhigh) | 83.89 |
| 38thof 298 | GPT-5.4 Pro | Score (%) (xhigh) | 83.33 |
| 44thof 298 | DeepSeek V4.1 Flash 1017 | Score (%) (max) | 78.06 |
| 46thof 298 | Gemini 3.1 Pro Preview | Score (%) | 77.08 |
| 48thof 298 | dots3-note-prev | Score (%) (max) | 76.81 |
| 50thof 298 | Claude Opus 4.7 | Score (%) (max) | 75.83 |
| 52ndof 298 | GPT-5.4 | Score (%) (xhigh) | 73.95 |
| 54thof 298 | DeepSeek V4.1 Flash | Score (%) (max) | 72.92 |
| 56thof 298 | Claude Opus 4.8 | Score (%) (high) | 72.08 |
| 57thof 298 | Gemini 3.5 Flash | Score (%) (high) | 72.08 |
| 61stof 298 | Claude Opus 4.6 | Score (%) (high) | 69.17 |
| 72ndof 298 | Grok 4.6 | Score (%) (xhigh) | 67.08 |
| 74thof 298 | GLM 5.3 Flash | Score (%) (max) | 65.83 |
| 81stof 298 | DeepSeek V4 Flash (0731) | Score (%) (max) | 61.39 |
| 82ndof 298 | Grok 4.7 | Score (%) (xhigh) | 61.39 |
| 84thof 298 | DeepSeek V4 Pro 0813 | Score (%) (max) | 61.25 |
| 85thof 298 | Claude Haiku 5.5 | Score (%) (xhigh) | 61.25 |
| 87thof 298 | Claude Sonnet 4.6 | Score (%) (high) | 60.42 |
| 88thof 298 | Kimi K3 | Score (%) (max) | 60.42 |
| 89thof 298 | Gemini 3.6 Flash | Score (%) (high) | 60.42 |
| 91stof 298 | GPT-5.6 Luna | Score (%) (max) | 59.58 |
| 94thof 298 | GPT 6 Luna | Score (%) (max) | 59.31 |
| 99thof 298 | Muse Spark Open | Score (%) (xhigh) | 57.5 |
| 104thof 298 | GPT-5.2 Pro | Score (%) (high) | 54.16 |
| 108thof 298 | Grok 4.5 | Score (%) (high) | 52.64 |
| 115thof 298 | Gemini 3 Deep Think | Score (%) | 45.14 |
| 119thof 298 | Qwen3.8 27B | Score (%) (xhigh) | 42.36 |
| 122ndof 298 | Inkling Small | Score (%) (xhigh) | 40.14 |
| 126thof 298 | Inkling | Score (%) | 36.53 |
| 135thof 298 | gemini-3-pro-preview | Score (%) | 31.11 |
| 151stof 298 | GLM 5.2 | Score (%) | 22.78 |
| 154thof 298 | GPT-5.4 mini | Score (%) (xhigh) | 18.9 |
| 156thof 298 | GPT-5 Pro | Score (%) | 18.33 |
| 158thof 298 | GPT 5.1 Thinking | Score (%) (high) | 17.64 |
| 169thof 298 | Kimi K2.5 | Score (%) | 11.81 |
| 170thof 298 | Gemini 3.5 Flash-Lite | Score (%) (high) | 10.28 |
| 171stof 298 | GPT-5 | Score (%) (high) | 9.86 |
| 176thof 298 | Claude Opus 4 | Score (%) | 8.61 |
| 183rdof 298 | o3 | Score (%) (high) | 6.53 |
| 187thof 298 | o4-mini | Score (%) (high) | 6.11 |
| 188thof 298 | Claude Sonnet 4 | Score (%) | 5.93 |
| 190thof 298 | GPT-5.4 nano | Score (%) (xhigh) | 5.69 |
| 194thof 298 | o3-pro | Score (%) (high) | 4.86 |
| 196thof 298 | MiniMax M2.5 | Score (%) | 4.86 |
| 197thof 298 | GLM-5 | Score (%) | 4.86 |
| 201stof 298 | GPT-5 mini | Score (%) (high) | 4.44 |
| 206thof 298 | DeepSeek v3.2 | Score (%) | 4.03 |
| 209thof 298 | Gemini 2.5 Pro | Score (%) | 3.75 |
| 210thof 298 | Claude Sonnet 4.5 | Score (%) | 3.75 |
| 215thof 298 | o3-mini | Score (%) (high) | 2.99 |
| 222ndof 298 | GPT-5 nano | Score (%) (high) | 2.61 |
| 240thof 298 | Nemotron 3.5 Super VL 120B A12B | Score (%) (medium) | 1.94 |
| 241stof 298 | Gemini 2.5 Flash | Score (%) | 1.69 |
| 251stof 298 | Gemini 2.0 Flash | Score (%) | 1.3 |
| 253rdof 298 | Codex mini | Score (%) | 1.27 |
| 256thof 298 | Qwen3-235B-A22B-Instruct-2507 | Score (%) | 1.25 |
| 257thof 298 | Claude Haiku 4.5 | Score (%) | 1.25 |
| 261stof 298 | DeepSeek-R1 (0528) | Score (%) | 1.12 |
| 266thof 298 | o1-mini | Score (%) | 83 |
| 276thof 298 | GPT-4.1 fine-tuned | Score (%) | 42 |
| 277thof 298 | Grok 3 Mini | Score (%) (low) | 42 |
| 281stof 298 | GPT-4o (2024-11-20) | Score (%) | 0 |
| 282ndof 298 | GPT-4o-mini (2024-07-18) | Score (%) | 0 |
| 285thof 298 | GPT-4.1 nano | Score (%) | 0 |
| 286thof 298 | GPT-4.1 mini | Score (%) | 0 |
| 289thof 298 | Grok 3 | Score (%) | 0 |
| 290thof 298 | Magistral Small 2506 | Score (%) | 0 |
Read from the board on 2026-10-31