{"api_version":"codesota.v1","data":{"announcement":"https://www.llamaindex.ai/blog/parsebench","area_id":"vision-documents","candidate_local_pipelines":["Docling","Dots OCR","Qwen VLM"],"canonical_task_ids":["document-parsing","table-recognition","document-layout-analysis"],"code_license":"Apache-2.0","code_revision":"9af4102e26b51f234ab5f037e19fec05c0696ec6","dataset":"llamaindex/ParseBench","dataset_license_metadata":"apache-2.0","dataset_revision":"2805a1d940f95a203e0ae4b88be9934f7765b3fc","dimensions":["tables","charts","content faithfulness","semantic formatting","visual grounding"],"domains":["enterprise"],"id":"parsebench","integration_requirements":["Pin evaluator, dataset files and split/group identifiers","Adapt each local parser to the required structured output and preserve native coordinates","Keep all five dimensions separate; do not infer document parsing from line CER","Record local model revision, parser configuration, failures, latency and reserved GPU cost","Keep upstream vendor-reported scores separate from CodeSOTA runs"],"methods":["document-parsing"],"modalities":["document-image","pdf"],"name":"ParseBench","note":"Candidate pipeline names from upstream documentation; exact models, weights licences, local hosting and adapter compatibility still require validation. No CodeSOTA ParseBench scores yet.","retrieved_at":"2026-10-18T17:34:11.255724+00:00","runnable":false,"safety_properties":["faithfulness","grounding"],"scoring":"Default evaluator is deterministic; optional judge-based chart normalization must remain disabled or a separate protocol.","source":"https://github.com/run-llama/ParseBench","status":"adapter_required"}}