Skip to main content
The following listings apply to Let’s Go, Pay-As-You-Go, and Business SaaS users of Unstructured Pipelines and API. For dedicated instance and in-VPC deployments, see AI providers.
Use this page to find the model IDs you can set for each workflow node, grouped by node and provider. This information is accurate as of the time of publication and may change as models are added or deprecated. This page does not list deprecated models. If a workflow specifies a deprecated model that has a replacement, Unstructured uses the replacement model instead. If a deprecated model has no replacement, Unstructured rejects the workflow.

Transform nodes

Partitioner (VLM and Auto)

The VLM and Auto partitioning strategies use the same models.

Partitioner (Multimedia)

The Multimedia strategy uses a transcription model for audio and a vision model for video scenes. It is available for Unstructured Pay-As-You-Go and Business SaaS accounts. To choose the vision model, set video_vlm_provider and video_vlm_model to a pair listed under Partitioner (VLM and Auto). If you set neither, Unstructured uses its default vision model.

Enrichment nodes

Contextual Chunking

Specify the model in provider/model format, for example, anthropic/claude-sonnet-4-6.

Chunk by Table Merging

Embedder

Extract (LLM)

Generative OCR

Image Description

Named Entity Recognition

Table Description

Table to HTML