Reducto r-1: A single-pass document parsing model at 1 cent per page

Reducto released r-1, a new document parsing model that replaces its multi-stage agentic OCR pipeline with a single full-page pass. The model is available in preview through the Parse API on V3, and Reducto claims it reduces errors by 20% compared to its legacy pipeline while costing 1 cent per page, all in.

What r-1 changes in document parsing

Legacy Parse ran OCR, layout detection, and post-processing as separate stages, with optional agentic vision-language passes layered on top. Each extra model call added latency and cost. r-1 folds text, tables, figures, layout, reading order, formatting, and grounding into a single pass. Every block returns with page-relative bounding boxes that tie content back to its position on the page. The consolidation is the real product claim: teams working on financial statements, insurance claims, or contracts often route files across several providers and bolt on post-processing to reach usable accuracy. r-1 targets that orchestration cost, not only raw character accuracy.

Pricing drops from the legacy 3 to 6 cents per page to a flat 1 cent, with no feature multipliers or credit costs layered on to reach high accuracy. Reducto frames this as part of a broader move to flat product rate cards.

Deployment options include multi-tenant cloud, customer VPC, on premises, and air-gapped installs, with SOC 2 Type II attestation and HIPAA processing on higher tiers.

Why the consolidation matters for document-heavy workflows

For builders processing regulated documents, the single-pass architecture reduces both latency and pipeline complexity. Instead of orchestrating OCR, table extraction, and layout analysis across separate services, r-1 handles all of it in one API call. The flat pricing also simplifies cost forecasting: no surprise multipliers for handwriting, tables, or high-resolution scans.

Reducto calls out edge cases that matter in production: dense tables, unusual layouts, low-quality scans, watermarked content, and documents that follow no predictable template. A dropped strikethrough can invert a contract clause, and a misread table can hand an agent the wrong figure. r-1's native handling of formatting like bold, underlines, and strikethroughs is designed to catch those details.

What builders should know before adopting r-1

r-1 requires the V3 Parse API. A Parse request that omits settings.model still runs on legacy Parse, so nothing breaks silently. New pipelines created in Studio default to r-1. Teams migrating an existing configuration should check the r-1 configuration compatibility page first, since some legacy settings are ignored or unsupported.

Agentic processing is not gone. Workflows needing custom prompts or advanced chart extraction still route those pages through the agentic pipeline, which Reducto augments onto the r-1 result, adding latency. Reducto has also signalled two upcoming pieces: r-1 mini for speed and cost-sensitive workloads, and automatic per-page routing that picks the right model for each page.

There are no open weights and no self-hosting option. r-1 runs only through Reducto's hosted Parse API. Organizations moving off another parser can request up to $5,000 in credits for a side-by-side comparison.

Caveats to keep in mind

The 20% error reduction is measured relative to Reducto's own prior pipeline, not a third-party baseline. The head-to-head comparisons against hyperscalers like Amazon Textract and Azure Document Intelligence are vendor-run, with no public evaluation harness or dataset released alongside the announcement. r-1 is in preview, so capabilities and pricing may change before general availability.