Codesota · Computer Vision · Document Parsing · ParseBenchTasks/Computer Vision/Document Parsing
Document Parsing · benchmark dataset · 2026 · EN

ParseBench: A Document Parsing Benchmark for AI Agents.

LlamaIndex's 2026 benchmark for document parsing evaluated against agent-facing semantic correctness. ~2,078 human-verified pages from ~1,211 enterprise documents (insurance, finance, government) with 169,011 rule-based tests across five dimensions: tables (GTRM: GriTS + TableRecordMatch), charts (ChartDataPointMatch), content faithfulness, semantic formatting, and visual grounding (element pass rate). Purely rule-based — no LLM-as-judge. Overall score is the unweighted mean across the five dimensions.

Paper ↗Download datasetSubmit a result ↵
§ 01 · Leaderboard

Best published scores.

14 results indexed across 1 metric. Shaded row marks current SOTA; ties broken by submission date.

Primary
accuracy · higher is better
accuracy· primary
14 rows
#ModelOrgSubmittedPaper / codeaccuracy
01LlamaParse AgenticLlamaIndexApr 2026blog-post84.90
02LlamaParse Cost EffectiveLlamaIndexApr 2026blog-post71.90
03Gemini 3 FlashAPIGoogleApr 2026blog-post71
04ReductoReductoApr 2026blog-post67.80
05Qwen3-VL-4BOSSAlibaba QwenApr 2026blog-post62
06Azure Document IntelligenceMicrosoftApr 2026blog-post59.60
07ExtendExtendApr 2026blog-post55.80
08Dots OCR 1.5OSSRedNote HILabApr 2026blog-post55.80
09DoclingOSSIBM ResearchApr 2026blog-post50.60
10Google Cloud Document AIGoogle CloudApr 2026blog-post50.40
11AWS TextractAmazon Web ServicesApr 2026blog-post47.90
12GPT-5-miniOpenAIApr 2026blog-post46.80
13LandingAILandingAIApr 2026blog-post45.20
14Anthropic Haiku 4.5AnthropicApr 2026blog-post45.20
Fig 2 · Rows sorted by score within each metric. Shaded row marks SOTA. Dates reflect model or paper release where available, otherwise the date Codesota accessed the source.
§ 03 · Progress

1 steps
of state of the art.

Each row below marks a model that broke the previous record on accuracy. Intermediate submissions are kept in the leaderboard above; only SOTA-setting entries are re-listed here.

Higher scores win. Each subsequent entry improved upon the previous best.

SOTA line · accuracy
  1. Apr 14, 2026LlamaParse AgenticLlamaIndex84.90
Fig 3 · SOTA-setting models only. 1 entries span Apr 2026 → Apr 2026.
§ 06 · Contribute

Have a score that beats
this table?

Submit a checkpoint and a reproduction script. We will run it, publish the score, and — if it takes the top — annotate the step on the progress chart with your name.

Submit a result ↵Read submission guide
What a submission needs
  • 01A public checkpoint or API endpoint
  • 02A reproduction script with frozen commit + seed
  • 03Declared evaluation environment (Python, deps)
  • 04One row per metric declared by this dataset
  • 05A contact so we can follow up on discrepancies