GLM-OCR is a multimodal OCR model for complex document understanding, built on the GLM-V encoder–decoder architecture. It introduces Multi-Token Prediction (MTP) loss and stable full-task reinforcement learning to improve training efficiency, recognition accuracy, and generalization.
MiniCPM-OCR-1.0
openbmb·Instruct·released 21 Oct 2026
ollama run hf.co/mradermacher/MiniCPM-OCR-1.0-GGUF:Q4_K_MMiniCPM-OCR-1.0 is an instruct-tuned OCR and document AI model from openbmb with 1.3B parameters. It takes image and text inputs and returns text, with a context length of 262,144 tokens. It is released under the Apache 2.0 licence, which allows commercial use, and it works with English and Chinese. The model has no other sizes in its family listed on this page.
How much VRAM does MiniCPM-OCR-1.0 need?
At FP16 / BF16 MiniCPM-OCR-1.0 needs about 3.4 GB of GPU memory.
| Quantisation | Total |
|---|---|
| 4-bitRecommended default for local use | ≈1.1 GB |
| 8-bitLossless in practice, twice the size of Q4 | ≈1.4 GB |
| FP16 / BF16recommendedFull precision, as released | ≈2.2 GB |
Estimates for the weights plus runtime overhead; activations vary with resolution and batch size. Pick a GPU above and the fit column checks every row against it. Weights marked “file” use the real GGUF download size; “est.” rows are computed from the parameter count.
Run MiniCPM-OCR-1.0 locally
--quantQ4_K_Mdefault; pick your GPU to tailor itmradermacher/MiniCPM-OCR-1.0-GGUF
ollama run hf.co/mradermacher/MiniCPM-OCR-1.0-GGUF:Q4_K_M llama-server -hf mradermacher/MiniCPM-OCR-1.0-GGUF:Q4_K_MDownloads the GGUF on first run. Installed through llama.app? The same flags work with `llama serve`.
lms get mradermacher/MiniCPM-OCR-1.0-GGUF@Q4_K_MThen load it from the LM Studio app or with `lms load`.
vllm serve openbmb/MiniCPM-OCR-1.0 vllm serve openbmb/MiniCPM-OCR-1.0-AWQservices:
vllm:
image: vllm/vllm-openai:latest
command: openbmb/MiniCPM-OCR-1.0
ports:
- "8000:8000"
ipc: host
volumes:
- ~/.cache/huggingface:/root/.cache/huggingface
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: all
capabilities: [gpu]API on port 8000.
Quantised versions and downloads
| Repository | Format | Quant | Size |
|---|---|---|---|
| mradermacher/MiniCPM-OCR-1.0-GGUF | GGUF | Q2_K | 0.4 GB |
| mradermacher/MiniCPM-OCR-1.0-GGUF | GGUF | Q3_K_S | 0.4 GB |
| mradermacher/MiniCPM-OCR-1.0-GGUF | GGUF | Q3_K_M | 0.5 GB |
| mradermacher/MiniCPM-OCR-1.0-GGUF | GGUF | Q3_K_L | 0.5 GB |
| mradermacher/MiniCPM-OCR-1.0-GGUF | GGUF | IQ4_XS | 0.5 GB |
| mradermacher/MiniCPM-OCR-1.0-GGUF | GGUF | Q4_K_S | 0.5 GB |
| mradermacher/MiniCPM-OCR-1.0-GGUF | GGUF | Q4_K_M | 0.5 GB |
| mradermacher/MiniCPM-OCR-1.0-GGUF | GGUF | Q5_K_S | 0.6 GB |
| mradermacher/MiniCPM-OCR-1.0-GGUF | GGUF | Q5_K_M | 0.6 GB |
| mradermacher/MiniCPM-OCR-1.0-GGUF | GGUF | Q6_K | 0.6 GB |
About MiniCPM-OCR-1.0
MiniCPM-OCR 1.0 (1.3B) is the first OCR-specific model in the MiniCPM-V/o family. It focuses on text-centric document understanding — text, tables, formulas, charts, seals, and structured information extraction — and balances quality against speed on real-world captures. • 🏆 SOTA document parsing: 93.68 on Real5-OmniDocBench, a benchmark for real-world • document parsing, with leading results on OmniDocBench v1.6, Wild-OmniDocBench, and • olmOCR-Bench. Handles Chinese and English documents and PDFs, including scans, curved pages, • photographs, skewed pages, and recaptured documents. • ⚡ Efficient inference. With DSpark decoding acceleration • (BF16, GGUF) end-to-end • processing efficiency improves by 55%. Throughput is about 1.49× PaddleOCR-VL-1.6 and • 3.62× MinerU2.5 Pro. • 🔍 General recognition and information extraction: 54.7 on OCRBench v2, plus key • information extraction (KIE) for downstream automation. • 🛠️ Easy to deploy and develop: vLLM, SGLang, llama.cpp, and Ollama from day one, with • quantized variants and multi-chip builds. Supports further fine-tuning, and ships a pip SDK, • a full API, and agent workflows. Online demos: China · • Global. Among the open-source OCR models compared, MiniCPM-OCR 1.0 reaches 1.2 pages/s and the highest equal-weight average score across OmniDocBench v1.6, Real5-OmniDocBench, Wild-OmniDocBench, and olmOCR-Bench.
Full model card on Hugging Face →Alternatives to MiniCPM-OCR-1.0
All alternatives →📦 Model layout. This repository hosts HunyuanOCR-1.5 checkpoint at the root (target base weights). The DFlash speculative-decoding draft lives under dflash/, and the previous HunyuanOCR-1.0 is archived under v1.0/ (load it with subfolder="v1.0", or download the v1.0/ directory directly).
Best OCR model . LightOnOCR-2-1B is LightOn's flagship OCR model, refined with RLVR training for maximum accuracy. We recommend this variant for most OCR tasks.
OvisOCR3 is a unified 0.8B multimodal model for OCR and document understanding. It integrates document parsing, general OCR, key information extraction, document figure captioning, and cross-page table reconstruction into a single compact model.
TeleOCR is a lightweight (~1.2B parameters), open-source Vision-Language Model designed specifically for document parsing.
MinerU2.5-Pro is our latest document parsing model (PDF-to-Markdown) that establishes a new industry standard. By focusing entirely on data engineering without altering the original 1.2B-parameter architecture, it delivers exceptional results across the board:
MiniCPM-OCR-1.0 FAQ
- How much VRAM does MiniCPM-OCR-1.0 need?
- About 3.4 GB at FP16 / BF16.
- Can MiniCPM-OCR-1.0 run on the RTX 4060?
- Yes. It fits at FP16 / BF16 (about 2.2 GB). It fits comfortably on the Apple M4 (16 GB), Radeon RX 6700 XT, RTX 3060 12GB.
- Can I use MiniCPM-OCR-1.0 commercially?
- Yes. It is released under Apache 2.0. Permissive: use, modify and sell freely; keep the notice and licence text.
- How do I run MiniCPM-OCR-1.0 with Ollama or llama.cpp?
- With Ollama: ollama run hf.co/mradermacher/MiniCPM-OCR-1.0-GGUF:Q4_K_M. With llama.cpp: llama-server -hf mradermacher/MiniCPM-OCR-1.0-GGUF:Q4_K_M, which starts an OpenAI-compatible server.
Rate this model
Data from Hugging Face, the model card and public leaderboards. Memory and speed figures are estimates; real numbers depend on the runtime, context length and drivers. Always check the licence yourself before commercial use.
0 comments