Sakhanda Wire
NVDA $153.63 -1.61% MSFT $506.84 -2.15% GOOGL $280.27 -1.22% META $619.77 -1.06% AMZN $256.65 -0.43%
← Back to the news

Reducto unveils r-1 for single-pass document parsing

Reducto announced r-1 on September 1, 2026, as the first model in a new document-parsing family. The system replaces the company’s earlier multi-stage agentic OCR workflow with a single pass over a full page. It is available in preview through the V3 hosted Parse API and is enabled with a configuration setting. Reducto has not released open weights or a local checkpoint for self-hosting.

According to Reducto, r-1 combines recognition of digital text, scans and handwriting with layout analysis, tables, figures, reading order, formatting and page-relative bounding boxes. The documentation also lists support for merged table cells, nested headers, columns, sidebars, figure descriptions and formatting such as headings, lists, bold, underlining and strikethroughs. Agentic processing remains available for workflows requiring custom prompts or advanced chart extraction, although it adds latency.

Reducto reports a 20% lower error rate than its own legacy agentic pipelines and says r-1 performed better than selected hyperscaler products and large language models in internal tests. The company prices it at 1 cent per page, compared with 3–6 cents for its previous agentic models. The comparisons were vendor-run, with no public dataset or evaluation harness released. New Studio pipelines default to r-1, while existing API requests remain on the legacy system unless the model is specified.

Written by Sakhanda Wire AI from reporting by MarkTechPost. Facts, quotes and full context are in the original article below.

Originally published by MarkTechPost on

Read the original on MarkTechPost ↗

The reporting and the image belong to MarkTechPost and are credited and linked above; the text on this page is Sakhanda Wire's own summary.

← Back to the news

More stories

All the latest news