Calibration
AA-Omniscience Accuracy → ARC-AGI-3
ARC-AGI-3 is estimated from AA-Omniscience Accuracy with a offset logistic curve fitted on 16 models measured on both: y = 0.0168 + (0.7350 − 0.0168) / (1 + exp(−123.71·(x − 0.6128))), R² = 0.98, cross-validated error 3.4 pp. It is used for 115 estimates.
| Estimated model | AA-Omniscience Accuracy | ARC-AGI-3 | Source |
|---|---|---|---|
| A.X K2 | 18.6% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Apodex 1.1 | 31.7% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Apodex 1.1 Mini | 31.7% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Celeris-1 | 11.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Claude 3 Haiku | 17.6% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Claude 4 Sonnet | 22.7% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Claude Fable 5 | 65.4% | 73.1% | estimated ± 3.4 pp, medium confidence |
| Claude Fable 5.1 | 67.2% | 73.5% | estimated ± 3.4 pp, medium confidence |
| Claude Opus 4.5 Thinking | 46.6% | 1.7% | estimated ± 3.4 pp, high confidence |
| Claude Opus 4.6 (Adaptive) | 47.0% | 1.7% | estimated ± 3.4 pp, high confidence |
| Claude Opus 4.7 | 44.7% | 1.7% | estimated ± 3.4 pp, high confidence |
| Claude Opus 5.5 | 66.2% | 73.3% | estimated ± 3.4 pp, medium confidence |
| Claude Sonnet 5 | 40.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Claude Sonnet 5.5 | 54.0% | 1.7% | estimated ± 3.4 pp, high confidence |
| Command A+ | 8.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| DeepSeek-R1 | 30.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| DeepSeek V3 0324 | 24.3% | 1.7% | estimated ± 3.4 pp, medium confidence |
| DeepSeek V3.1 | 23.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| DeepSeek V3.1 (Reasoning) | 29.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| DeepSeek V3.2 | 24.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Exaone 4.0 1.2B | 5.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Exaone 4.0 32B | 10.6% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Gemini 2.5 Flash | 26.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Gemini 3.5 Flash-Lite | 29.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Gemini 3.6 Flash | 50.0% | 1.7% | estimated ± 3.4 pp, high confidence |
| Gemini 3.7 Flash | 55.3% | 1.7% | estimated ± 3.4 pp, high confidence |
| Gemini 3 Flash | 45.8% | 1.7% | estimated ± 3.4 pp, high confidence |
| Gemini 3 Pro | 55.8% | 1.8% | estimated ± 3.4 pp, high confidence |
| Gemini 4 Argon | 49.9% | 1.7% | estimated ± 3.4 pp, high confidence |
| Gemma 3 27B | 13.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Gemma 4 26B A4B | 19.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GLM-4.5-Air | 16.3% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GLM-4.6 | 21.4% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GLM-5.1 | 23.7% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GLM-5.3 | 33.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GLM-5-Turbo | 28.4% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GLM-5V-Turbo | 29.3% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GPT-4o | 19.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GPT-5.1 | 37.7% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GPT-5.1-Codex | 39.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GPT-5.1-Codex-Max | 39.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GPT-5.2-Codex | 41.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GPT-5.3 Codex | 52.9% | 1.7% | estimated ± 3.4 pp, high confidence |
| GPT-5 (high) | 40.3% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GPT-5 (medium) | 39.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GPT-OSS 120B | 21.8% | 1.7% | estimated ± 3.4 pp, medium confidence |
| GPT-OSS 20B | 16.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Granite-4.0-350M | 3.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Granite-4.0-H-1B | 5.2% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Granite-4.0-H-350M | 3.8% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Grok 4 | 40.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Grok 4.1 Fast | 17.2% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Grok 4.1 Fast (Reasoning) | 25.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Grok 4.7 | 47.4% | 1.7% | estimated ± 3.4 pp, high confidence |
| Grok 4 Fast (Reasoning) | 22.8% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Grok Code Fast 1 | 23.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Hy3 | 32.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| K-Exaone | 16.4% | 1.7% | estimated ± 3.4 pp, medium confidence |
| K-EXAONE 2.0 | 13.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Kimi K2 | 27.4% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Kimi K2.7 Code | 39.6% | 1.7% | estimated ± 3.4 pp, medium confidence |
| LFM2.5-2.6B | 4.4% | 1.7% | estimated ± 3.4 pp, medium confidence |
| LFM2.5-8B-A1B | 9.4% | 1.7% | estimated ± 3.4 pp, medium confidence |
| LFM2.5-VL-1.6B-Extract | 5.8% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Ling 3.0 Flash VL | 14.4% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Ling 3.0 Tiny | 8.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Ling 3.1 Flash | 29.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Llama 3.1 405B | 23.2% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Llama 4 Maverick | 24.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Llama 4 Scout | 15.2% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Mercury 2.5 | 22.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| MiMo-V2.5-Pro | 22.4% | 1.7% | estimated ± 3.4 pp, medium confidence |
| MiMo-V2.6-Flash | 27.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| MiMo-V2.6-Pro | 34.8% | 1.7% | estimated ± 3.4 pp, medium confidence |
| MiMo-V2-Omni | 19.3% | 1.7% | estimated ± 3.4 pp, medium confidence |
| MiMo-V2-Pro | 26.6% | 1.7% | estimated ± 3.4 pp, medium confidence |
| MiniCPM5-2B | 8.4% | 1.7% | estimated ± 3.4 pp, medium confidence |
| MiniMax M2.7 | 26.8% | 1.7% | estimated ± 3.4 pp, medium confidence |
| MiniMax M3 | 16.7% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Mistral Large 2 | 19.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Mistral Large 3 | 25.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Mistral Large 4 | 25.8% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Mistral Medium 3 | 18.3% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Mistral Medium 3.5 128B | 24.7% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Mistral Small 4 | 21.7% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Mistral Small 4 (Reasoning) | 21.7% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Muse Glimmer 30B | 27.0% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Muse Spark | 49.6% | 1.7% | estimated ± 3.4 pp, high confidence |
| Muse Spark 1.1 | 52.1% | 1.7% | estimated ± 3.4 pp, high confidence |
| Muse Spark 1.2 | 45.4% | 1.7% | estimated ± 3.4 pp, high confidence |
| Muse Spark 1.3 | 43.6% | 1.7% | estimated ± 3.4 pp, high confidence |
| Nemotron 3 Nano 30B | 17.3% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Nemotron 3 Super 100B | 24.3% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Nemotron Ultra 253B | 20.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| North Mini Code | 18.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Nova Pro | 16.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| o3 | 38.6% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Phi-4 | 14.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Quasar 438B | 15.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Qwen3.5 397B (Reasoning) | 24.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Qwen 3.6 Max (preview) | 37.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Qwen3.8 Max Preview | 31.7% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Qwen3 Max | 24.4% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Qwen3-Omni-30B-A3B-Instruct | 14.3% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Qwen3-Omni-30B-A3B-Thinking | 14.6% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Sarvam 105B | 17.6% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Sarvam 30B | 12.6% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Solar Pro 2 | 16.1% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Solar Pro 3 | 18.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Solar Pro 4 | 18.9% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Step 3.7 Flash | 25.8% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Step 5 Preview | 41.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Trinity-Large-Preview | 22.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Trinity-Large-Thinking | 22.5% | 1.7% | estimated ± 3.4 pp, medium confidence |
| Ultravox v0.6 Llama 3.3 70B | 19.0% | 1.7% | estimated ± 3.4 pp, medium confidence |