Skip to content
Trending storyMonitoring

LlamaIndex introduces ParseBench-2 document parsing benchmark and OCRia-1.5 models

1 article1 sourcesince Nov 12Last article

Overview

AISummary of 1 article

LlamaIndex introduces ParseBench-2, a document parsing benchmark with more than 4,300 pages, double the size of its first version, and a new forms dimension.

The benchmark scores parsing errors across more than 330 error types using severity-tuned metrics. LlamaIndex says these metrics track human judgment better than deterministic or LLM-as-a-judge benchmarks.

The company also releases OCRia-1.5, a new family of document parsing models that powers all LlamaParse modes.

Written by AI from the articles below · updated Nov 12, 12:11 PM ET

Check the sources:

Latest developmentNov 12, 11:46 AM ET
LlamaIndex introduces ParseBench-2 document parsing benchmark and OCRia-1.5 models

Article timeline

The articles in this story. Times are ET.

Nov 12
  1. Jerry LiuX
    LlamaIndex introduces ParseBench-2 document parsing benchmark and OCRia-1.5 models

    AILlamaIndex introduces ParseBench-2, a document parsing benchmark with 4.3k+ pages, double the first version, and a new forms dimension. It scores parsing errors across 330+ error types with severity-tuned metrics, which the company says track human judgment better than deterministic or LLM-as-a-judge benchmarks. LlamaIndex also releases OCRia-1.5, a new family of document parsing models that powers all LlamaParse modes.

    Video from @jerryjliu0's post

Heat Trend

Current comparable heat
4
Peak comparable heat
10Nov 12, 1:00 PM ET
Change in past 24 hours
-51%

The trend compares only continuously observed participants, so its scope may be smaller than the current heat total. Hover or tap for hourly heat; use arrow keys with a keyboard.