{"name":"digital-pages","title":"Digital pages","keyPoints":["LiteParse, with no model, reads 20 of 20 values on the two digital pages in 0.01 s a page, at no cost. Mistral OCR 4.1 also reads 20 of 20, in 9 to 13 s a page, for $5.00 per 1,000 pages.","Poppler's embedded text reads the same 20 of 20 at no cost, but as spaced lines: LiteParse writes the fee table as a table, under its heading.","Claude Haiku 5.5 and GPT-6 Luna read 20 of 20 at low effort and 19 of 20 at medium: at medium each writes a row's second $750.00 under a wrong column."],"tables":[{"heading":"Two digital pages read from their PDF: correct values, time, tokens and cost","columns":["Reader","Effort","Correct","Time, fee table","Time, two tables","Tokens in","Tokens out","Per 1,000 pages"],"rows":[["Embedded text (Poppler, no model)","none","20 of 20","0.06 s","0.05 s","none","none","$0"],["LiteParse, its OCR off (no model)","none","20 of 20","0.01 s","0.01 s","none","none","$0"],["LiteParse, its OCR on (no model)","none","20 of 20","0.74 s","0.75 s","none","none","$0"],["Claude Haiku 5.5","low","20 of 20","12.6 s","7.0 s","3,076","3,149","$1.88"],["Claude Haiku 5.5","medium","19 of 20","14.6 s","9.7 s","3,076","3,774","$2.19"],["GPT-6 Luna","low","20 of 20","18.9 s","48.4 s","3,664","1,866","$1.30"],["GPT-6 Luna","medium","19 of 20","18.6 s","17.5 s","3,664","2,274","$1.50"],["Mistral OCR 4.1 with page annotation, from the page image","none","20 of 20","9.4 s","12.9 s","billed by the page","billed by the page","$5.00"]],"note":"Two digital pages of test sets A and B, 10 checked values each: a fee table, and two tables side by side. One reading a page, one page a request, from a development computer on Oct 9 2026 (LiteParse: Oct 25 2026, the median of three), not on sheaf.us. Tokens: the mean of the two pages, reasoning included. Medium is each model's default effort. Embedded text is the PDF's text as Poppler gives it, its layout kept by spacing: lines, not tables. LiteParse 2.17.0 reads the same text layer and rebuilds its headings and tables by rules: its OCR off as Digital pages · LiteParse runs it, on as the parse engine LiteParse."}],"scoringMethod":"20 values on two digital pages of the published test sets A and B, each checked by a person, counted when the reading states the value where it belongs. The grader is Claude Opus 5.5, reading the text alone. Time is one reading of one page, one page a request, on a development computer.","log":[{"date":"261025","line":"New board. LiteParse reads 20 of 20 values on two digital pages in 0.01 s a page at no cost; Mistral OCR 4.1 20 of 20 at $5.00 for 1,000 pages."}],"updatedAt":"2026-10-25T18:59:31.610Z"}