None
None
Long Story Video Skill
Turn your idea into 10s–10min video—AI writes scripts, prompts, and generates footage automatically.
Ads Video Skill
Generate professional ads and sales videos—AI auto-generates scripts, prompts, and footage.
3D Science Explainer Video Skill
Convert scientific concepts into stunning 3D explain animations
Feedback
freeTrialImage.bannerPity
freeTrialImage.upgradeUnlock
- ✓freeTrialImage.benefitHd
- ✓freeTrialImage.benefitWatermark
- ✓freeTrialImage.benefitUnlimited
xiaomi ocr 0
Turn scanned pages, tables, formulas and handwriting into clean structured text with Xiaomi OCR 0 — a tiny 0.8B vision-language model you can try free.
All Tools
Discover our comprehensive AI-powered animation toolkit
MiniMax H3
MiniMax H3 AI Video Generator
Seedance 2.5
The Future of AI Video Is Here.
Seedance 2.0
The Future of AI Video Is Here.
Veo3.1
Create Stunning Videos with Veo3.1
FLUX 3 Video Generator
Kling 3.0
Next-Gen AI Video Generator
Grok Video Generator
Create Videos from Text or Images with AI
MiniMax H3 video generator
MiniMax H3 AI Video Generator
Inside Xiaomi OCR 0: One Engine, Every Page Type
A single 0.8B checkpoint handles page parsing, OCR-driven Q&A and key-field extraction — replacing a whole stack of narrow utilities with one download.
- All-in-One 0.8B CheckpointPage parsing, OCR-focused visual Q&A and field extraction share one lightweight set of weights.
- Benchmark LeaderIt ranks first on Real5-OmniDocBench with 95.24 and takes second on OmniDocBench v1.6 as well as the Wild variant.
- Reliable on Messy ScansWarped, skewed, photographed or photocopied sheets still parse cleanly — not just crisp digital PDFs.
Run Xiaomi OCR 0 in Three Steps
From a raw page image to tidy structured text in three quick moves.
Core Capabilities of Xiaomi OCR 0
Compact size, resilient parsing, handwritten-script support and openly released weights define what this 0.8B model does well.
Page Parsing
Covers ordinary layouts along with scanned, bent, tilted and re-photographed sheets.
Tables and Equations
Equation recognition hits 97.84 on UniMER-Test, and table structure scores 92.27 on PubTabNet.
Handwritten Scripts
Reaches 73 on Chronicles-OCR for mature handwriting, matching Seed 2.0 Pro.
OCR-Focused Visual Q&A
Averages 83.2 over five visual question answering suites, edging past a rival twice its size.
Structured Field Capture
Returns organized fields from receipts and forms rather than flat transcription.
Open Weights, Local Runs
Apache-2.0 checkpoints ship with GGUF and BF16 versions for llama.cpp, Ollama and Transformers.
Xiaomi OCR 0: Questions Answered
Straight answers on what this compact model does, how it scores and where you can run it.
What exactly is this model?
It is a 0.8B vision-language checkpoint tuned for document reading: it parses pages and answers OCR-style questions in a single pass.
Where does it rank on benchmarks?
First on Real5-OmniDocBench, second on OmniDocBench v1.6 and on Wild-OmniDocBench.
Does it handle handwritten text?
Yes — clerical, cursive, running and regular calligraphy all work, scoring 73 on Chronicles-OCR.
What went into its training?
Around 170M OCR-heavy samples, Q-Mask text anchoring, continued pretraining and mixed-task reinforcement learning.
Can it run offline on my machine?
Yes. Apache-2.0 weights are public, and community GGUF and BF16 builds work with llama.cpp, Ollama and Transformers.
Is transcription all it does?
No — it also answers questions about a page and returns key fields as structured output.
Put Xiaomi OCR 0 to Work on Your Files
Upload one page image and get back clean text, tables and structured fields in moments.