Document Structure Analysis on StructBench
4.77Tree Edit Distance (TED)LingoEDU
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| LingoEDUType=Specialized, Backbone=Qwen3-4B, Deployment=Local2025.12 | 4.77 | 0.496 | 0.17 | 1.2 | |
| Claude-4-SonnetType=General LLM, Access=API2025.12 | 5.08 | 0.4315 | 7.09 | — | |
| OpenAI o3Type=General LLM, Access=API2025.12 | 5.51 | 0.2863 | 4.17 | — | |
| Gemini-2.5-ProType=General LLM, Access=API2025.12 | 5.61 | 0.3266 | 4.02 | — | |
| Gemini-2.5-FlashType=General LLM, Access=API2025.12 | 5.82 | 0.2782 | 0.99 | — | |
| OpenAI o4-miniType=General LLM, Access=API2025.12 | 5.87 | 0.3266 | 2.28 | — | |
| GPT-4oType=General LLM, Access=API2025.12 | 6.22 | 0.2903 | 5.21 | — | |
| DeepSeek-R1Type=General LLM, Access=API2025.12 | 6.26 | 0.3065 | 1.14 | — | |
| DeepSeek-V3Type=General LLM, Access=API2025.12 | 6.32 | 0.3347 | 0.3 | — | |
| GPT-4.1Type=General LLM, Access=API2025.12 | 6.35 | 0.379 | 4.17 | — | |
| Claude-3.7-SonnetType=General LLM, Access=API2025.12 | 6.65 | 0.3508 | 7.09 | — | |
| Qwen3-32BType=General LLM, Deployment=Local2025.12 | 9.49 | 0.249 | 0.26 | 10.17 | |
| Qwen3-235BType=General LLM2025.12 | 9.93 | 0.1789 | 0.11 | — | |
| FirecrawlType=Parser API2025.12 | 16.81 | — | 0.17 | — | |
| Jina-ReaderType=Parser API2025.12 | 17.04 | — | 0.1 | — |