Mistral Ships OCR 4: Self-Hosted, Structure-Aware Document AI in a Single Container
Mistral AI·high signal
On June 23 Mistral released OCR 4, a document-intelligence model that returns bounding boxes, block classification, and inline confidence scores across 170 languages and runs fully self-hosted in one container — aimed at regulated enterprises that cannot route documents to cloud APIs. It tops OlmOCRBench at 85.20 with a 72% average human-preference win rate over rival OCR systems, priced at $4 per 1,000 pages ($2 via the Batch API). For builders running RAG over private corpora, it is a notable on-prem option for the ingestion layer.