Source: LlamaParse Tiers and Pricing Docs

LlamaIndex documentation for LlamaParse v2: the Fast, Cost-effective, Agentic and Agentic Plus parse tiers, dated version pinning, caching, and per-page credit pricing as of 2026-11.

Contents
  • What: official product documentation (tiers guide and pricing page)
  • Author/publisher: LlamaIndex
  • Date: living documents; read 2026-11-14
  • Original: Parse tiers, Pricing

Summary

LlamaParse is LlamaIndex’s hosted document parser. Version 2 replaces many knobs with four tiers that trade cost for layout reasoning, from a no-model text extractor up to a separate high-accuracy pipeline for dense financial documents. Requests must name both a tier and a version, and production users are told to pin a dated version.

Key claims

  • Fast: no AI model over the page; for high-volume plain text; no chart extraction. 1 credit/page.
  • Cost-effective: runs on the OCRia-1 model; clean Markdown with table preservation for text-heavy documents with simple tables. 3 credits/page.
  • Agentic: same base, more layout reasoning; for scans, multi-column layouts, real tables, charts and diagrams; a custom prompt switches to a general-purpose model. 10 credits/page.
  • Agentic Plus: a separate pipeline for dense financial reports and complex tables, with chart parsing on by default. 25 credits/page on the pricing page (the FAQ still listed 45 when read; treat the pricing page as authoritative).
  • Layout extraction adds 3 credits/page; spreadsheets cost 1 credit per sheet.
  • Credits cost $1.50 per 1,000 pay-as-you-go and $1.25 on Pro (as of 2026-11).
  • tier and version are required; pin a dated version (for example 2026-01-08) in production so new releases do not change output.
  • Re-parsing the same file within 48 hours is free (cached). A Cost Optimizer routes each page of mixed-complexity documents to an appropriate tier.

Feeds

Document processing

Suggest changes →