Skip to content
NewA calmer, faster ZeroHour

OCR costs fall fivefold to under $130 per million pages

Original title: OCR models now digitize a million pages for under $130 as costs drop fivefold in 18 months

The Decoderby Jonathan Kemper ()1 min read1
AI summary · x-ai/grok-4.7Summaries by Vani AI

Hugging Face's Daniel van Strien timed eight OCR models on one Nvidia RTX PRO 6000 rented at $2.75 an hour.

Van Strien compared eight freely available OCR models on 850 pages from old natural history books, the 1771 Encyclopaedia Britannica, and typed and printed archival documents. At the same image size, RolmOCR from April 2025 cost about 51 cents per 1,000 pages, while LightOnOCR-3-0.8B from October 2026 came in just under 10 cents. He said runtime changes saved about as much as newer models, and the newest model matched the oldest on book-page accuracy, though some models hallucinated or repeated text.

Unconfirmed: Costs and accuracy are from one comparison of 850 pages, not a broad production benchmark.

Key facts
  • A million pages cost $65 to $130 in GPU rental, depending on the material.
  • The five models between the oldest and newest were almost on par with each other.
  • Savings came from even page feeding, smaller scans, software updates, and a faster GPU.
  • Some models hallucinated text on blank pages or repeated the same content.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

1 min read at the-decoder.com

Pulling text out of scanned books and records has gotten a lot cheaper with freely available AI models. The article OCR models now digitize a million pages for under $130 as costs drop fivefold in 18 months appeared first on The Decoder.

The Decoder

By Jonathan Kemper · the-decoder.com

Read the full article at The Decoder

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.