Calibration
AA-MMMU-Pro → OfficeQA Pro
OfficeQA Pro is estimated from AA-MMMU-Pro with a offset logistic curve fitted on 8 models measured on both: y = 0.5032 + (0.6772 − 0.5032) / (1 + exp(−200.00·(x − 0.8034))), R² = 0.66, cross-validated error 7.2 pp. It is used for 96 estimates.
| Estimated model | AA-MMMU-Pro | OfficeQA Pro | Source |
|---|---|---|---|
| Apodex 1.1 | 79.2% | 51.9% | estimated ± 7.2 pp, medium confidence |
| Apodex 1.1 Mini | 79.2% | 51.9% | estimated ± 7.2 pp, medium confidence |
| Claude 3 Haiku | 30.8% | 50.3% | estimated ± 7.2 pp, low confidence |
| Claude 4.1 Opus Thinking | 67.9% | 50.3% | estimated ± 7.2 pp, low confidence |
| Claude 4 Sonnet | 62.4% | 50.3% | estimated ± 7.2 pp, low confidence |
| Claude Opus 4.5 | 71.2% | 50.3% | estimated ± 7.2 pp, low confidence |
| Claude Opus 4.5 Thinking | 74.0% | 50.3% | estimated ± 7.2 pp, low confidence |
| Claude Opus 4.6 | 72.5% | 50.3% | estimated ± 7.2 pp, low confidence |
| Claude Opus 4.6 (Adaptive) | 75.4% | 50.3% | estimated ± 7.2 pp, low confidence |
| Claude Opus 4.7 | 76.4% | 50.3% | estimated ± 7.2 pp, medium confidence |
| Claude Sonnet 4.6 | 70.6% | 50.3% | estimated ± 7.2 pp, low confidence |
| Claude Sonnet 5 | 77.3% | 50.4% | estimated ± 7.2 pp, medium confidence |
| Command A+ | 63.2% | 50.3% | estimated ± 7.2 pp, low confidence |
| DeepSeek V4.1 Flash | 77.0% | 50.3% | estimated ± 7.2 pp, medium confidence |
| Gemini 1.5 Pro | 55.0% | 50.3% | estimated ± 7.2 pp, low confidence |
| Gemini 2.5 Flash | 65.5% | 50.3% | estimated ± 7.2 pp, low confidence |
| Gemini 2.5 Pro | 74.9% | 50.3% | estimated ± 7.2 pp, low confidence |
| Gemini 3.1 Pro | 82.4% | 67.4% | estimated ± 7.2 pp, medium confidence |
| Gemini 3.5 Flash | 84.3% | 67.7% | estimated ± 7.2 pp, medium confidence |
| Gemini 3.5 Flash-Lite | 79.0% | 51.4% | estimated ± 7.2 pp, medium confidence |
| Gemini 3.6 Flash | 83.2% | 67.7% | estimated ± 7.2 pp, medium confidence |
| Gemini 3.7 Flash | 85.5% | 67.7% | estimated ± 7.2 pp, medium confidence |
| Gemini 3.8 Flash | 85.6% | 67.7% | estimated ± 7.2 pp, medium confidence |
| Gemini 3 Flash | 78.6% | 50.8% | estimated ± 7.2 pp, medium confidence |
| Gemini 3 Pro | 80.2% | 57.8% | estimated ± 7.2 pp, medium confidence |
| Gemma 3 27B | 48.0% | 50.3% | estimated ± 7.2 pp, low confidence |
| Gemma 4 12B | 69.7% | 50.3% | estimated ± 7.2 pp, low confidence |
| Gemma 4 26B A4B | 69.2% | 50.3% | estimated ± 7.2 pp, low confidence |
| Gemma 4 31B | 73.4% | 50.3% | estimated ± 7.2 pp, low confidence |
| Gemma 4 E2B | 44.6% | 50.3% | estimated ± 7.2 pp, low confidence |
| Gemma 4 E4B | 51.4% | 50.3% | estimated ± 7.2 pp, low confidence |
| GLM-5V-Turbo | 72.8% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-4.1 | 61.2% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-4.1 mini | 58.7% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-4.1 nano | 40.1% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-4o mini | 41.5% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-5.1 | 75.5% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-5.1-Codex | 72.5% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-5.1-Codex-Max | 72.5% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-5.2-Codex | 76.3% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-5.3 Codex | 78.5% | 50.7% | estimated ± 7.2 pp, medium confidence |
| GPT-5.4 mini | 73.3% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-5.4 nano | 65.4% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-5.6 Luna | 78.6% | 50.8% | estimated ± 7.2 pp, medium confidence |
| GPT-5.6 Sol | 83.4% | 67.7% | estimated ± 7.2 pp, medium confidence |
| GPT-5.6 Terra | 80.7% | 62.0% | estimated ± 7.2 pp, medium confidence |
| GPT-5 (high) | 74.2% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-5 (medium) | 74.3% | 50.3% | estimated ± 7.2 pp, low confidence |
| GPT-6.1 Sol | 86.0% | 67.7% | estimated ± 7.2 pp, medium confidence |
| GPT-6 Astra | 86.9% | 67.7% | estimated ± 7.2 pp, medium confidence |
| GPT-6 Luna | 79.7% | 54.1% | estimated ± 7.2 pp, medium confidence |
| GPT-6 Sol | 82.9% | 67.6% | estimated ± 7.2 pp, medium confidence |
| Grok 4 | 68.8% | 50.3% | estimated ± 7.2 pp, low confidence |
| Grok 4.1 Fast | 48.4% | 50.3% | estimated ± 7.2 pp, low confidence |
| Grok 4.1 Fast (Reasoning) | 63.3% | 50.3% | estimated ± 7.2 pp, low confidence |
| Grok 4.3 | 78.1% | 50.5% | estimated ± 7.2 pp, medium confidence |
| Grok 4.5 | 80.4% | 59.6% | estimated ± 7.2 pp, medium confidence |
| Grok 4 Fast (Reasoning) | 61.8% | 50.3% | estimated ± 7.2 pp, low confidence |
| Inkling | 73.5% | 50.3% | estimated ± 7.2 pp, low confidence |
| Inkling-Small | 74.0% | 50.3% | estimated ± 7.2 pp, low confidence |
| Kimi K2.6 | 79.4% | 52.6% | estimated ± 7.2 pp, medium confidence |
| Kimi K2.5 | 75.4% | 50.3% | estimated ± 7.2 pp, low confidence |
| Kimi K2.5 (Reasoning) | 75.4% | 50.3% | estimated ± 7.2 pp, low confidence |
| LFM2.5-VL-1.6B-Extract | 26.5% | 50.3% | estimated ± 7.2 pp, low confidence |
| Ling 3.0 Flash VL | 79.0% | 51.4% | estimated ± 7.2 pp, medium confidence |
| Llama 4 Maverick | 62.1% | 50.3% | estimated ± 7.2 pp, low confidence |
| Llama 4 Scout | 52.9% | 50.3% | estimated ± 7.2 pp, low confidence |
| MiMo-V2.6-Flash | 73.1% | 50.3% | estimated ± 7.2 pp, low confidence |
| MiMo-V2-Omni | 69.9% | 50.3% | estimated ± 7.2 pp, low confidence |
| Mistral Large 3 | 55.7% | 50.3% | estimated ± 7.2 pp, low confidence |
| Mistral Large 4 | 76.4% | 50.3% | estimated ± 7.2 pp, medium confidence |
| Mistral Medium 3 | 53.0% | 50.3% | estimated ± 7.2 pp, low confidence |
| Mistral Medium 3.5 128B | 64.9% | 50.3% | estimated ± 7.2 pp, low confidence |
| Mistral Small 4 | 56.8% | 50.3% | estimated ± 7.2 pp, low confidence |
| Mistral Small 4 (Reasoning) | 56.8% | 50.3% | estimated ± 7.2 pp, low confidence |
| Muse Glimmer 30B | 74.3% | 50.3% | estimated ± 7.2 pp, low confidence |
| Muse Spark | 80.5% | 60.4% | estimated ± 7.2 pp, medium confidence |
| Nemotron 3 Nano Omni 30B A3B | 53.2% | 50.3% | estimated ± 7.2 pp, low confidence |
| Nova Pro | 44.3% | 50.3% | estimated ± 7.2 pp, low confidence |
| o3 | 70.1% | 50.3% | estimated ± 7.2 pp, low confidence |
| Phi-4 Multimodal Instruct | 14.5% | 50.3% | estimated ± 7.2 pp, low confidence |
| Qwen3.5-122B-A10B | 75.0% | 50.3% | estimated ± 7.2 pp, low confidence |
| Qwen3.5-27B | 75.0% | 50.3% | estimated ± 7.2 pp, low confidence |
| Qwen3.5-35B-A3B | 72.7% | 50.3% | estimated ± 7.2 pp, low confidence |
| Qwen3.5 397B | 52.7% | 50.3% | estimated ± 7.2 pp, low confidence |
| Qwen3.5 397B (Reasoning) | 52.7% | 50.3% | estimated ± 7.2 pp, low confidence |
| Qwen3.6-27B | 74.6% | 50.3% | estimated ± 7.2 pp, low confidence |
| Qwen3.6-35B-A3B | 75.0% | 50.3% | estimated ± 7.2 pp, low confidence |
| Qwen3.6 Plus | 78.0% | 50.5% | estimated ± 7.2 pp, medium confidence |
| Qwen3.7 Plus | 80.5% | 60.4% | estimated ± 7.2 pp, medium confidence |
| Qwen3.8-27B | 76.3% | 50.3% | estimated ± 7.2 pp, low confidence |
| Qwen3.8-Flash-Next | 79.8% | 54.7% | estimated ± 7.2 pp, medium confidence |
| Qwen3.8 Max Preview | 82.8% | 67.6% | estimated ± 7.2 pp, medium confidence |
| Qwen3-Omni-30B-A3B-Instruct | 55.5% | 50.3% | estimated ± 7.2 pp, low confidence |
| Qwen3-Omni-30B-A3B-Thinking | 60.2% | 50.3% | estimated ± 7.2 pp, low confidence |
| Step 3.7 Flash | 75.3% | 50.3% | estimated ± 7.2 pp, low confidence |