Model card
Mistral OCR 3.
MistralapiVision-Language Model1 current SOTA
Latest Mistral OCR (Dec 2025). 74% win rate vs OCR 2. Claims 94.9% accuracy. Markdown + HTML table output. $1/1000 pages with batch API.
§ 01 · Benchmarks
Every benchmark Mistral OCR 3 has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | OmniDocBench | Computer Vision · Document Parsing | reading-order | 91.6% | #1 | — | source ↗ |
| 02 | OmniDocBench | Computer Vision · Document Parsing | formula-edit-distance | 0.2% | #2 | — | source ↗ |
| 03 | OmniDocBench | Computer Vision · Document Parsing | text-edit-distance | 0.1% | #2 | — | source ↗ |
| 04 | OmniDocBench | Computer Vision · Document Parsing | table-teds | 70.9% | #3 | — | source ↗ |
| 05 | olmOCR-Bench | Computer Vision · Document Parsing | pass-rate | 78.0% | #11 | — | source ↗ |
| 06 | OmniDocBench | Computer Vision · Document Parsing | composite | 79.8% | #26 | — | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 02 · Strengths by area
Where Mistral OCR 3 actually performs.
§ 04 · Related models
Other Mistral models scored on Codesota.
§ 05 · Sources & freshness
Where these numbers come from.
codesota-verified
4
results
mistral-announcement
1
result
CodeSOTA (verified)
1
result
5 of 6 rows marked verified.