Data Extraction
Subcategory of Utility. 91 models scored.
Model Leaderboard
All models ranked by their Data Extraction subcategory score.
| # | Model | Data Extraction | Utility | Overall |
|---|---|---|---|---|
| 1 | Claude Opus 4.6 (Reasoning) | 100.00% | 98.93% | 95.02% |
| 2 | Gemini 3.1 Pro (Preview) | 100.00% | 99.91% | 94.37% |
| 3 | Claude Sonnet 4.6 (Reasoning) | 100.00% | 97.88% | 93.66% |
| 4 | GPT-5 Mini | 100.00% | 98.39% | 92.62% |
| 5 | GPT-5.1 | 100.00% | 95.33% | 92.54% |
| 6 | Claude Opus 4.6 | 100.00% | 90.72% | 92.35% |
| 7 | Qwen 3.5 397B A17B | 100.00% | 97.50% | 91.73% |
| 8 | Qwen 3.5 122B | 100.00% | 96.36% | 91.53% |
| 9 | Z.AI GLM 5 | 100.00% | 94.11% | 91.23% |
| 10 | Claude Sonnet 4.6 | 100.00% | 88.52% | 91.15% |
| 11 | MoonshotAI: Kimi K2.5 | 100.00% | 96.63% | 91.04% |
| 12 | Qwen 3.5 27B | 100.00% | 95.67% | 90.85% |
| 13 | ByteDance Seed 1.6 | 100.00% | 90.83% | 90.70% |
| 14 | Gemini 3 Flash (Preview, Reasoning) | 100.00% | 97.20% | 90.50% |
| 15 | o4 Mini High | 100.00% | 98.67% | 90.29% |
| 16 | GPT-5.2 | 100.00% | 96.22% | 90.26% |
| 17 | Claude Opus 4.5 | 100.00% | 89.84% | 89.69% |
| 18 | Grok 4.1 Fast | 100.00% | 84.12% | 89.55% |
| 19 | Aion 2.0 | 100.00% | 90.91% | 89.21% |
| 20 | Z.AI GLM 4.6 | 100.00% | 88.58% | 89.11% |
| 21 | Gemini 3 Pro (Preview) | 100.00% | 96.14% | 88.79% |
| 22 | Claude Sonnet 4 | 100.00% | 84.02% | 88.72% |
| 23 | Minimax M2.5 | 100.00% | 90.42% | 88.71% |
| 24 | Z.AI GLM 4.7 | 100.00% | 94.31% | 88.69% |
| 25 | GPT-4.1 | 100.00% | 90.57% | 88.68% |
| 26 | o4 Mini | 100.00% | 96.31% | 88.35% |
| 27 | Grok 4 | 100.00% | 89.67% | 88.12% |
| 28 | Claude Sonnet 4.5 | 100.00% | 83.78% | 88.03% |
| 29 | Qwen 3.5 35B | 100.00% | 96.42% | 88.00% |
| 30 | Claude Opus 4 | 100.00% | 88.81% | 87.69% |
| 31 | Qwen 3.5 Flash | 100.00% | 96.11% | 86.38% |
| 32 | Z.AI GLM 4.5 | 100.00% | 79.19% | 86.27% |
| 33 | Grok 4 Fast | 100.00% | 76.76% | 86.15% |
| 34 | Qwen 3.5 Plus (2026-02-15) | 100.00% | 86.65% | 85.96% |
| 35 | Gemini 2.5 Flash Lite (Reasoning) | 100.00% | 89.63% | 85.75% |
| 36 | Mistral Large 3 | 100.00% | 84.91% | 85.43% |
| 37 | GPT-4o, May 13th (temp=0) | 100.00% | 83.13% | 85.36% |
| 38 | Gemini 3 Flash (Preview) | 100.00% | 86.39% | 85.35% |
| 39 | Claude Haiku 4.5 | 100.00% | 72.48% | 85.14% |
| 40 | DeepSeek-V2 Chat | 100.00% | 83.82% | 84.83% |
| 41 | Claude 3.5 Sonnet | 100.00% | 76.75% | 84.24% |
| 42 | GPT-4o, May 13th (temp=1) | 100.00% | 80.69% | 83.80% |
| 43 | Stealth: Aurora Alpha | 100.00% | 92.59% | 83.79% |
| 44 | Claude 3.5 Haiku | 100.00% | 82.57% | 83.73% |
| 45 | DeepSeek V3 (2024-12-26) | 100.00% | 81.87% | 83.68% |
| 46 | Claude 3.7 Sonnet | 100.00% | 62.54% | 83.39% |
| 47 | GPT-4.1 Mini | 100.00% | 82.30% | 83.20% |
| 48 | GPT-4o, Aug. 6th (temp=1) | 100.00% | 82.44% | 82.62% |
| 49 | GPT-5 Nano | 100.00% | 93.91% | 82.60% |
| 50 | GPT-4o, Aug. 6th (temp=0) | 100.00% | 82.11% | 82.45% |
| 51 | Mistral Large 2 | 100.00% | 69.19% | 82.41% |
| 52 | DeepSeek V3.1 | 100.00% | 76.65% | 82.39% |
| 53 | DeepSeek V3.2 | 100.00% | 81.58% | 82.25% |
| 54 | DeepSeek V3 (2025-03-24) | 100.00% | 80.62% | 81.99% |
| 55 | Gemini 2.5 Flash Lite | 100.00% | 80.14% | 81.08% |
| 56 | Mistral Large | 100.00% | 73.04% | 80.15% |
| 57 | Writer: Palmyra X5 | 100.00% | 79.71% | 79.57% |
| 58 | GPT-4o Mini (temp=1) | 100.00% | 82.16% | 79.08% |
| 59 | Mistral Small 3.2 24B | 100.00% | 73.17% | 78.60% |
| 60 | Gemma 3 12B | 100.00% | 79.28% | 78.41% |
| 61 | Llama 3.1 70B | 100.00% | 81.03% | 78.40% |
| 62 | GPT-4o Mini (temp=0) | 100.00% | 81.43% | 78.29% |
| 63 | Gemma 3 27B | 100.00% | 76.82% | 77.85% |
| 64 | Mistral Medium 3.1 | 100.00% | 80.13% | 77.83% |
| 65 | Qwen 2.5 72B | 100.00% | 76.43% | 75.46% |
| 66 | Llama 3.1 Nemotron 70B | 100.00% | 88.31% | 74.70% |
| 67 | ByteDance Seed 1.6 Flash | 100.00% | 84.16% | 73.27% |
| 68 | Mistral Small Creative | 100.00% | 76.28% | 73.27% |
| 69 | Hermes 3 70B | 100.00% | 61.15% | 72.57% |
| 70 | Ministral 3 14B | 100.00% | 79.03% | 72.54% |
| 71 | Ministral 3 8B | 100.00% | 74.43% | 71.76% |
| 72 | Claude 3 Haiku | 100.00% | 68.47% | 71.19% |
| 73 | WizardLM 2 8x22b | 100.00% | 67.14% | 71.07% |
| 74 | Arcee AI: Trinity Mini | 100.00% | 59.94% | 70.90% |
| 75 | Gemma 3 4B | 100.00% | 60.30% | 68.57% |
| 76 | Ministral 3 3B | 100.00% | 72.38% | 67.22% |
| 77 | Mistral NeMO | 100.00% | 51.55% | 65.04% |
| 78 | LFM2 24B | 100.00% | 69.48% | 58.77% |
| 79 | GPT-5 | 95.00% | 93.53% | 91.93% |
| 80 | Gemini 2.5 Pro | 95.00% | 92.18% | 88.53% |
| 81 | Z.AI GLM 4.7 Flash | 95.00% | 88.98% | 84.82% |
| 82 | Llama 3.1 8B | 95.00% | 74.82% | 63.37% |
| 83 | Rocinante 12B | 95.00% | 48.47% | 54.55% |
| 84 | Gemini 2.5 Flash (Reasoning) | 90.00% | 82.25% | 86.51% |
| 85 | Hermes 3 405B | 90.00% | 69.02% | 82.86% |
| 86 | Arcee AI: Trinity Large (Preview) | 85.00% | 60.74% | 73.33% |
| 87 | Ministral 8B | 85.00% | 46.82% | 64.87% |
| 88 | Ministral 3B | 85.00% | 49.17% | 61.29% |
| 89 | Cohere Command R+ (Aug. 2024) | 70.00% | 59.51% | 69.03% |
| 90 | Gemini 2.5 Flash | 50.00% | 61.45% | 80.60% |
| 91 | GPT-4.1 Nano | 50.00% | 68.45% | 71.94% |