Alibaba Cloud: Qwen-VL-OCR

OCR model for extracting structured text from documents and screenshots

qwen-vl-ocr

AvailableAlibaba CloudChat

Input / Output: $0.07 / $0.16
Context: 30K
Max output: 8K
Modalities: VisionChat

About Qwen-VL-OCR

Qwen-VL-OCR is Alibaba's text recognition model, built on Qwen3-VL, for pulling text and structured data out of images such as scanned documents, tables, receipts and screenshots. It ships built-in tasks for plain recognition, key information extraction, table parsing, document parsing and formula recognition, and can correct image rotation.

Where it works well

Built-in task modes cover text recognition, high-precision recognition with positions, table parsing to HTML, and formula recognition to LaTeX.
Key information extraction turns receipts, certificates and forms into JSON using the fields you define.
Image rotation correction helps with photographed pages that are skewed.

When to choose another model

Alibaba lists Qwen3.5-OCR as the newer model with better document parsing and certificate extraction.
Output is text, JSON, HTML or LaTeX only, never an edited or annotated image.
It reads images, not free-form conversation, and is not meant for open-ended reasoning.

Getting started

Create API key
Create a key in Console, then use it with every model on the platform.

Send your first request

Copy the example for your language and run it against the endpoint.

Text to textMessages API

POST/v1/messages

API format:

curl https://api.tokenlab.sh/v1/messages \
  -H "Content-Type: application/json" \
  -H "x-api-key: sk-xxx" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "qwen-vl-ocr",
    "max_tokens": 1024,
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Default

per 1M tokens

Official price: Input $0.07 / Output $0.16 / Text Input $0.07 / Text Output $0.16
Verified price: —
Discount: —

	Official priceper 1M tokens	Verified priceper 1M tokens	Discount
Default	Input $0.07 / Output $0.16 / Text Input $0.07 / Text Output $0.16	—	—

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours

Requests
Success rate
P95 latency
Total tokens

Model performance

Last updated Nov 1, 2026, 6:33 AM UTC

Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Qwen-VL-OCR in Console with a prompt ready to edit or send.

Help me try qwen-vl-ocr with a short message at /v1/messages. Show the reply, latency, and cost.

Use cases

Best for

Vision

Receipt and invoice capture
Extract merchant, date, line items and totals from receipts and invoices into JSON for accounting systems.
Table digitization
Convert a scanned or photographed table into HTML that a spreadsheet or database import can consume.
Formula extraction
Pull equations out of lecture slides or papers as LaTeX so they can be edited or typeset again.

Prompt examples

Extract all text from this image.

Read this receipt and return JSON with merchant, date, line items and total.

Convert the table in this scan to HTML.

This model has conditional pricing. Monthly token totals alone cannot produce a reliable estimate; use the detailed pricing for the request specification, cache, and applicable time window.

FAQ

What does Qwen-VL-OCR do?

It reads text from images and returns it as plain text, JSON, HTML or LaTeX, depending on the task. Its inputs are images such as scans, receipts and screenshots.

How does it differ from qwen3.5-ocr?

Qwen-VL-OCR is based on Qwen3-VL and has built-in task modes with rotation correction. Alibaba describes Qwen3.5-OCR as the latest architecture with enhanced document parsing, text localization and certificate extraction.

Which languages can it read?

Alibaba lists Chinese, English, Russian, French, Turkish, Italian, Portuguese, Hebrew and Arabic among the supported languages, and describes the support as multilingual rather than a closed list.

Can it extract structured fields?

Yes. The key information extraction task returns receipts, certificates and forms as JSON, using the fields you specify in the prompt, so downstream code can read the values directly.

Does it work for photos of documents?

Yes, and it can correct rotated images. Alibaba documents that images are scaled automatically within a stated pixel range, so very small or very large pictures are resized before recognition.

How much does Qwen-VL-OCR cost?

On TokenLab, Qwen-VL-OCR costs Input $0.07 / Output $0.16 per 1M tokens. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

What are the context window and output limit of Qwen-VL-OCR?

Qwen-VL-OCR accepts up to 30,000 tokens of context and returns up to 8,192 tokens in one response.

Which endpoint should Qwen-VL-OCR use?

Use https://api.tokenlab.sh/v1/messages for Qwen-VL-OCR. The request example below shows the matching code shape.

Which operations does Qwen-VL-OCR support?

Qwen-VL-OCR supports Text to text. Select an operation above to see its endpoint and request example.