Object-level Information Extraction on Insurance Claim Documents
90.42Object-level Accuracyxmemory
Evaluation results
| Method | Links | |
|---|---|---|
| xmemoryllm-judge-in-the-loop=true2026.04 | 90.42 | |
| Gemini 3.1 Pro preview2026.04 | 89.24 | |
| xmemory2026.04 | 86.61 | |
| OpenAI GPT-5.5high reasoning effort=true2026.04 | 83.98 | |
| Anthropic Sonnet 4.62026.04 | 83.56 | |
| Anthropic Opus 4.72026.04 | 83.31 | |
| OpenAI GPT-5.4advanced reasoning=true2026.04 | 81.69 | |
| OpenAI GPT-5.52026.04 | 81.02 | |
| OpenAI GPT-5.42026.04 | 79.24 |