Read receipts and handwriting 93% cheaper than Mistral OCR 4.1

GLM‑OCR reads receipt and handwriting photos into text, keeping amounts, dates and names as written. SIE is the open-source inference engine that serves this document model.

Get started Run in your cloud

GLM-OCR on SIE

Every amount as printed

Faded receipt rows: 1 Nila Bakar/Goreng 27,500; 1 Sop Gurame 87,000; 1 Teh Poci 25,000 — Full photo · NAVER Clova CC BY 4.0

Returned text

1 Nila Bakar/Goreng 27,500
1 Sop Gurame 87,000
1 Teh Poci 25,000

GPT‑6 Luna and Mistral OCR 4.1 change and drop receipt amounts

GPT-6 Luna (max reasoning)

Amounts changed in 3 of 5 runs

Returned text

1 Nila Bakar/Goreng 77.500
1 Sop Durian 37.000
1 Teh Poci 25.000

Mistral OCR 4.1

5 of 7 amounts missing

Receipt: Black Pepper Meatball Pas 79,000; Truffle Cream 77,000; Earl Grey Milk Tea 59,000; subtotal 215,000; service charge 12,900; tax 22,790; total 250,690 — Full photo · NAVER Clova CC BY 4.0

Returned text

79,000
1 BLACK PEPPER MEATBALL PAS
1 TRUFFLE CREAM
1 EARL GREY MILK TEA
SUBTTL
Service Charge 6%
TAX 10%
TOTAL
250,690

Not in the reply: 77,000 · 59,000 · 215,000 · 12,900 · 22,790

arXiv:2605.22413 Study of 10,656 receipts: AI models change line-item prices to make the total add up
From Recognition to Reasoning: Benchmarking and Enhancing MLLMs on Real-World Receipt Document Understanding

Yandi Wang, Libin Zhan, Ziwei Huang and others, Zhejiang University

“To force the sum of detail items to match the std_total, models occasionally alter the price of a line item”
21 May 2026
arXiv:2608.18586 CIKM 2026 Study of 13 AI models on handwriting: they may correct, complete or invent text instead of copying it
OmniHandwritingOCR: A Diagnostic Benchmark for Evaluating Multimodal LLMs in Handwritten OCR Scenarios

Zinuo Guo, Min Zhang and Bo Jiang, East China Normal University

“models may correct, complete, or fabricate content rather than transcribe what is visually present”
19 August 2026

93% cheaper than Mistral OCR 4.1 for every receipt and handwritten page

Run this eval

Photos read correctly

372 photos ↑

GPT-6.1 Sol (low reasoning)

GPT-6.1 Sol (low reasoning) 78.2%

Claude Sonnet 5.5

Claude Sonnet 5.5 78.0%

Claude Haiku 5.5

Claude Haiku 5.5 72.3%

GLM-OCR on SIE

GLM-OCR on SIE 71.0%

GPT-6 Luna (max reasoning)

GPT-6 Luna (max reasoning) 66.1%

Mistral OCR 4.1

Mistral OCR 4.1 62.1%

Cost per 1,000 photos

USD ↓

GLM-OCR on SIE

$0.25 GLM-OCR on SIE

Claude Haiku 5.5

$0.31 Claude Haiku 5.5

GPT-6 Luna (max reasoning)

$2.38 GPT-6 Luna (max reasoning)

Mistral OCR 4.1

Mistral OCR 4.1 $4.00

Claude Sonnet 5.5

Claude Sonnet 5.5 $6.40

GPT-6.1 Sol (low reasoning)

GPT-6.1 Sol (low reasoning) $6.63

Evaluation

Tested: Photo372: 200 receipt photos and 172 handwritten pages
Scored: Every line and every amount, date, name and number as written
Prices: Recorded API usage; Serverless and provider list prices

Send a photo, get its text back

from pathlib import Path
from sie_sdk import SIEClient
client = SIEClient(api_key="API keysk-sie-…")
image = {"data": Path(imagereceipt.jpgbrowse).read_bytes(), "format": "jpeg"}
result = client.extract(
    "modelzai-org/GLM-OCR",
    {"images": [image]},
)
print(result["entities"][0]["text"])

import { readFile } from 'node:fs/promises';
import { SIEClient } from '@superlinked/sie-sdk';

const client = new SIEClient(undefined, { apiKey: 'sk-sie-…' });

const image = await readFile("receipt.jpg");
const result = await client.extract(
  'zai-org/GLM-OCR',
  { images: [image] },
  { labels: [] },
);
console.log(result.entities[0].text);

images_bytes=$(base64 < 'receipt.jpg' | tr -d '\n')
curl https://api.superlinked.com/v1/extract/zai-org%2FGLM-OCR \
  -H "Authorization: Bearer sk-sie-…" \
  -H "Content-Type: application/json" \
  -d "{\"items\":[{\"images\":[{\"data\":\"$images_bytes\",\"format\":\"jpeg\"}]}]}"

Build the "OCR" capability into my app using the Superlinked Inference Engine (SIE).

Context
- SIE is an OpenAI-style inference API. Python SDK: `from sie_sdk import SIEClient`; TypeScript: `@superlinked/sie-sdk`.
- Base URL: https://api.superlinked.com (or my regional endpoint). Auth: Bearer key from env `SIE_API_KEY` (never hard-code it).
- Model: zai-org/GLM-OCR (SIE primitive: /extract). Keep the model id configurable.

Task
- Input: an uploaded image containing text.
- Behaviour: return the text visible in the image
- Call the selected SIE primitive once per request and map the response into your domain type.

Deliverables
- A typed client wrapper, an application-level function for this task, error handling for timeouts/empty input, and unit tests with a stubbed client.
- Wire it into my existing stack (ask me which framework if unclear) and add a short usage example.

Run

Output

2 Ayam Bakar 55,000
2 Nasi Putih 20,000
1 Nila Bakar/Goreng 27,500
1 Sop Gurame 87,000
1 Teh Poci 25,000

SUBTOTAL 214,500
Service 12,870
Pajak Restoran 22,737

TOTAL 250,107

Read a million photos for $250

Serverless API

We host the models; you call one API.

No idle GPUs, pay for what you use
Fits your stack: SDK, API, CLI, MCP
Self-host the same open-source engine
SOC 2 Type 2, US or EU data residency
Create an account to try Serverless

Get started

Self-hosted (Kubernetes)

Your data stays in your cloud, backed by our engineers.

Terraform modules for your cloud
Apache-2.0, same engine as Serverless
GPU workers scale to zero between jobs
Separate worker pools for each tenant

Agent prompt Deploy SIE to our AWS account with the superlinked/sie/aws Terraform module. Docs: superlinked.com/docs/deployment.

Deploy SIE to our Google Cloud project with the superlinked/sie/google Terraform module. Docs: superlinked.com/docs/deployment.

Deploy SIE to our Azure subscription with the superlinked/sie/azure Terraform module. Docs: superlinked.com/docs/deployment.

Deploy SIE to our Alibaba Cloud account with the superlinked/sie/alicloud Terraform module. Docs: superlinked.com/docs/deployment.

Talk to us Deploy guide

Run locally

The open-source server on your own machine.

NVIDIA GPU
One container, no cluster
Process your data fully offline
Same SDK and model IDs as Serverless

docker run --gpus all -p 8080:8080 ghcr.io/superlinked/sie-server:latest-cuda12-sglang-vision-extract

Quickstart

Switch to SIE in 5 minutes

import base64
from mistralai.client import Mistral

image = base64.b64encode(open("receipt.jpg", "rb").read()).decode()
client = Mistral(api_key="…")
result = client.ocr.process(
    model="mistral-ocr-4-1",
    document={"type": "image_url", "image_url": f"data:image/jpeg;base64,{image}"},
)
print(result.pages[0].markdown)

import base64
from openai import OpenAI

image = base64.b64encode(open("receipt.jpg", "rb").read()).decode()
reply = OpenAI().responses.create(
    model="gpt-6-luna",
    input=[{"role": "user", "content": [
        {"type": "input_text", "text": "Transcribe all readable text in the image."},
        {"type": "input_image", "image_url": f"data:image/jpeg;base64,{image}"},
    ]}],
)
print(reply.output_text)

import base64
import anthropic

image = base64.b64encode(open("receipt.jpg", "rb").read()).decode()
reply = anthropic.Anthropic().messages.create(
    model="claude-haiku-5-5",
    max_tokens=4096,
    messages=[{"role": "user", "content": [
        {"type": "text", "text": "Transcribe all readable text in the image."},
        {"type": "image", "source": {
            "type": "base64", "media_type": "image/jpeg", "data": image,
        }},
    ]}],
)
print(reply.content[0].text)

SIE

from sie_sdk import SIEClient

client = SIEClient(api_key="sk-sie-…")
image = {"data": open("receipt.jpg", "rb").read(), "format": "jpeg"}
result = client.extract("zai-org/GLM-OCR", {"images": [image]})
print(result["entities"][0]["text"])

from sie_sdk import SIEClient

client = SIEClient(api_key="sk-sie-…")
image = {"data": open("receipt.jpg", "rb").read(), "format": "jpeg"}
result = client.extract("zai-org/GLM-OCR", {"images": [image]})
print(result["entities"][0]["text"])

from sie_sdk import SIEClient

client = SIEClient(api_key="sk-sie-…")
image = {"data": open("receipt.jpg", "rb").read(), "format": "jpeg"}
result = client.extract("zai-org/GLM-OCR", {"images": [image]})
print(result["entities"][0]["text"])