What's the difference between Azure AI Foundry Content Understanding Standard and Pro and what does the roadmap look like for the future of Content Understanding and Pro?

Hello Xavier Hreben

Azure AI Foundry Content Understanding has two modes: Standard is the default, cost‑ and latency‑optimized option that extracts fields to a schema you define across all modalities (documents, images, audio, video, text), with optional data labeling for documents and new generative/classification fields; it’s ideal when you just need structured outputs (e.g., for RAG or analytics) without heavy inference or cross‑file logic. Pro adds multi‑step reasoning plus multi‑file inputs and lets you attach reference data at analyzer creation to perform linking, validation, enrichment, and derived decisions; today Pro applies to documents only (other modalities are noted as “coming soon”).

A concrete Pro example is a 3‑way match: ingest a PO, an invoice, and a goods‑receipt together, join lines, validate quantities and prices against a price master (reference data), compute exceptions, and emit an approve/hold decision work that goes beyond extraction to cross‑document inference and rule checking. You can build this in the Foundry UI by creating a hub‑based project → Content Understanding → +Create → Multi‑file task (Pro mode) and following the multi‑file quickstart.

For pricing, the public page currently lists Pro Field Extraction per‑token rates lower than Standard (e.g., Input $1.21/M tokens, Output $4.84/M vs Standard Input $2.75/M, Output $11/M), with Contextualization priced at $1.50/M tokens (Pro) vs $1/M (Standard); total cost depends on token volumes and whether Pro reduces retries or custom code.

Recent updates include the 2025‑05‑01‑preview API with Pro mode and Foundry support for creating Standard and Pro tasks, document classification & splitting, improved structure extraction (multi‑page tables, selection marks, barcodes), more file types (DOCX/XLSX/PPTX/MSG/EML/RTF/HTML/MD/XML), video chapters/whole‑video fields/face identification, and audio multilingual/expanded limits; the docs explicitly state Pro is documents‑only now with other content types “coming soon,” so track the What’s new page for Pro’s multi‑modal expansion and additional enterprise features as they light up.

Hope it helps!

Thank you

Hello Xavier Hreben

Do you get any chance to check above response?

Thank you

Hello Xavier Hreben

Just following up to see if you had a chance to review the above response.

Thank you!