Marketplace · Text & chat
glm-ocr:latest
Call glm-ocr:latest through one OpenAI-compatible API. It's served by 3 hosts on the Inpherio grid, each at a price it sets — your request goes to the cheapest healthy one, and fails over to the next if it's down.
At a glance.
vision ocr tools
from openai import OpenAI
client = OpenAI(base_url="https://inpherio.co.uk/v1",
api_key=INPHERIO_API_KEY)
reply = client.chat.completions.create(
model="glm-ocr:latest",
messages=[{"role": "user", "content": "Hello"}],
)glm-ocr:latest in detail.
GLM-OCR is a multimodal OCR model for complex document understanding, built on the GLM-V encoder–decoder architecture.
- Published by
- zai-org
- Parameters
- 1.3B
- Download size
- 2.7 GB (as packaged for Ollama)
- Quantisation
- F16
- Format
- gguf
- Architecture
- glm_ocr
- Licence
- MIT
- Task
- image to text
- Library
- transformers
- Languages
- zh, en, fr, es, ru, de, ja, ko
- First published
- 30 Jan 2026
- Last updated
- 11 Sep 2026
- Downloads (30 days)
- 1.8 million on Hugging Face
- Likes
- 2,205
glm_ocr image-text-to-text image-to-text zh en fr es ru de ja ko
The model card
Published by the model's authors on Hugging Face — shown as they wrote it, unedited by Inpherio.
GLM-OCR
👋 Join our WeChat and Discord community
📍 Use GLM-OCR's API
👉 GLM-OCR SDK Recommended
📖 Technical Report
Introduction
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. The model integrates the CogViT visual encoder pre-trained on large-scale image–text data, a lightweight cross-modal connector with efficient token downsampling, and a GLM-0.5B language decoder. Combined with a two-stage pipeline of layout analysis and parallel recognition based on PP-DocLayout-V3, GLM-OCR delivers robust and high-quality OCR performance across diverse document layouts.
Key Features
- State-of-the-Art Performance: Achieves a score of 94.62 on OmniDocBench V1.5, ranking #1 overall, and delivers state-of-the-art results across major document understanding benchmarks, including formula recognition, table recognition, and information extraction.
- Optimized for Real-World Scenarios: Designed and optimized for practical business use cases, maintaining robust performance on complex tables, code-heavy documents, seals, and other challenging real-world layouts.
- Efficient Inference: With only 0.9B parameters, GLM-OCR supports deployment via vLLM, SGLang, and Ollama, significantly reducing inference latency and compute cost, making it ideal for high-concurrency services and edge deployments.
- Easy to Use: Fully open-sourced and equipped with a comprehensive SDK and inference toolchain, offering simple installation, one-line invocation, and smooth integration into existing production pipelines.
Performance
- Document Parsing & Information Extraction
- Real-World Scenarios Performance
- Speed Test
For speed, we compared different OCR methods under identical hardware and testing conditions (single replica, single concurrency), evaluating their performance in parsing and exporting Markdown files from both image and PDF inputs. Results show GLM-OCR achieves a throughput of 1.86 pages/second for PDF documents and 0.67 images/second for images, significantly outperforming comparable models.
Usage
Official SDK
For document parsing tasks, we strongly recommend using our official SDK. Compared with model-only inference, the SDK integrates PP-DocLayoutV3 and provides a complete, easy-to-use pipeline for document parsing, including layout analysis and structured output generation. This significantly reduces the engineering overhead required to build end-to-end document intelligence systems.
Note that the SDK is currently designed for document parsing tasks only. For information extraction tasks, please refer to the following section and run inference directly with the model.
Serve GLM-OCR Locally
GLM-OCR supports deployment with the following frameworks. Feel free to try them out:
- SGLang — see cookbook
- vLLM — see recipes
- Transformers — see transformers docs
Read the rest of the model card
Prompt Limited
GLM-OCR currently supports two types of prompt scenarios:
- Document Parsing – extract raw content from documents. Supported tasks include:
{
"text": "Text Recognition:",
"formula": "Formula Recognition:",
"table": "Table Recognition:"
}
- Information Extraction – extract structured information from documents. Prompts must follow a strict JSON schema. For example, to extract personal ID information:
请按下列JSON格式输出图中信息:
{
"id_number": "",
"last_name": "",
"first_name": "",
"date_of_birth": "",
"address": {
"street": "",
"city": "",
"state": "",
"zip_code": ""
},
"dates": {
"issue_date": "",
"expiration_date": ""
},
"sex": ""
}
⚠️ Note: When using information extraction, the output must strictly adhere to the defined JSON schema to ensure downstream processing compatibility.
Acknowledgement
This project is inspired by the excellent work of the following projects and communities:
License
The GLM-OCR model is released under the MIT License.
The complete OCR pipeline integrates PP-DocLayoutV3 for document layout analysis, which is licensed under the Apache License 2.0. Users should comply with both licenses when using this project.
Citation
If you find GLM-OCR useful in your research, please cite our technical report:
@misc{duan2026glmocrtechnicalreport,
title={GLM-OCR Technical Report},
author={Shuaiqi Duan and Yadong Xue and Weihan Wang and Zhe Su and Huan Liu and Sheng Yang and Guobing Gan and Guo Wang and Zihan Wang and Shengdong Yan and Dexin Jin and Yuxuan Zhang and Guohong Wen and Yanfeng Wang and Yutao Zhang and Xiaohan Zhang and Wenyi Hong and Yukuo Cen and Da Yin and Bin Chen and Wenmeng Yu and Xiaotao Gu and Jie Tang},
year={2026},
eprint={2603.10910},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2603.10910},
}Sources: zai-org/GLM-OCR on Hugging Face · glm-ocr:latest on Ollama. Collected from public sources on 8 Nov 2026; the host serving it may run a different build or quantisation — the marketplace shows each host's own measurements.
Who's serving it.
Every host benchmarked this model on its own hardware before serving it. Prices are set by each host; health is scored by the platform from what it observes.
| Host | Tier | Price | Measured speed | Health | State | Rent |
|---|---|---|---|---|---|---|
| 7B-A1B (5G) | L1 | £0.044 in · £0.18 out | 26 tok/s · 41 ms first token | 82 | loads on request | £0.04/h |
| W11BOX | L1 | £0.03 in · £0.06 out | 18 tok/s · 94 ms first token | 77 | loads on request | — |
| GTX 1080 Ti (11G) | L1 | £0.043 in · £0.086 out | 23 tok/s · 18889 ms first token | 78 | loads on request | £0.60/h |
Live data, refreshed on every visit. How billing works.
Room on the grid for your machine.
3 machines serve glm-ocr:latest today. Every extra host is somewhere a request can fail over to when one drops, and spare capacity when demand spikes. The router weighs price, measured speed, health and whether the model is already loaded — so a host with a sharper price or quicker hardware wins the next call.
Set your own price — you keep 85% of what it earns, and new hosts start with 12 months fee-free (up to £250 of earnings). A payout fee of £2 + 1% per payout to a UK bank account, £2 + 2% to any other country applies when you're paid. Install the connector, load glm-ocr:latest, and the platform benchmarks it on your hardware before your first request.