Mistral AI·· 2026-06-23
Introducing Mistral OCR 4
Introducing Mistral OCR 4
AI summary
Mistral released OCR 4, returning bounding boxes, block-type classifications and per-page and per-word confidence alongside text extraction. It supports 170 languages and single-container self-hosting. Independent annotators preferred OCR 4 on average 72% of the time in blind evaluations of over 600 documents. It scored 85.20 on OlmOCRBench and 93.07 on OmniDocBench, though Mistral warns both benchmarks have known scoring limitations.
Selection record
Threshold 60Official, first-handFirst 62Second 71
AdmittedSum of both 133 ≥ twice the threshold 120
- Source tier
- Official, first-hand; this tier's threshold is 60
- Pre-filter
- passed：Mistral OCR 4模型发布，含AI技术细节
- Why it was chosen
- Bounding boxes, block classifications, confidence outputs, self-hosting and per-page pricing inform integration costs for RAG and document pipelines.
A model scores each item twice, independently, against one written standard, out of 100. An item is admitted only when the two scores add up to twice the threshold. The threshold is set per source tier.
Source: Mistral AI · mistral.ai