LlamaIndex releases OCRia-1, a 4B document parsing model

LlamaIndex released OCRia-1, a 4B-parameter VLM for document parsing, claiming SOTA on ParseBench (87.4) and leading three third-party benchmarks in accuracy, cost and latency, though some questioned the fairness of unified rendering in the olmOCR benchmark.

2026-10-22 ~ 2026-10-23 · 4 related posts

• LlamaIndex releases OCRia-1, a 4B document-parsing VLM claiming SOTA on ParseBench at 87.4 — llama_index · 2026-10-22
• LlamaIndex launches OCRia-1, claiming SOTA document parsing on accuracy, speed and cost — _nerdai_ · 2026-10-23
• LlamaIndex launches OCRia-1 VLM, but its olmOCR-bench methodology draws fire — IgorCarron · 2026-10-23

1 near-duplicate retellings: grilledCheeseFish