LlamaIndex Releases OCRia-1, Setting New SOTA in Document Parsing

LlamaIndex has released OCRia-1, a 4B vision-language model for document parsing, along with a 37-page technical report. It sets new SOTA on ParseBench with a score of 87.4 while decoding 2-4x faster.

2026-10-26 ~ 2026-10-26 · 2 related posts

• OCRia-1: 4B doc-parsing VLM hits 0.87 mean ParseBench, decodes 2-4x faster — NielsRogge · 2026-10-26
• LlamaIndex releases OCRia-1, topping ParseBench at 87.4 with 2-4x faster decoding — llama_index · 2026-10-26