GLM-4.5V · Answer Questions Using Images and Text
Ideal for describing images, extracting text from screenshots, or answering specific questions based on table screenshots. When asking a question, specify the region, field, or content you want to identify.
Model and Input/Output
- Request Model
- glm-4.5v
- Input
- Text and images
- Output
- Text
- Typical Tasks
- Image descriptions, OCR, and chart interpretation
How to Structure Input
Include text and image_url content blocks in the same user message. Use an image URL accessible to the service, and describe in text what you want the model to examine.
Making a Request
Send model and messages to POST /glm/chat/completions. For non-streaming responses, the text is located in choices[0].message.content.
Reading Responses Incrementally
After setting stream: true, concatenate choices[0].delta.content from SSE events in order, and save the complete response when the stream ends.
Usage and Pricing
Each million input tokens costs 4.906 Credits, and each million output tokens costs 4.731 Credits. Charges are calculated based on the public model specified in the request. When compatible routing is enabled, the platform can fulfill requests using compatible models in the same tier.
Frequently Asked Questions
What should I enter for model?
Enter glm-4.5v.
How do I submit an image together with a question?
Place a text object and an image_url object in the content array of the same user message, and put the image address in image_url.url.
How can I improve table screenshot recognition accuracy?
Upload clear screenshots in the correct orientation, and specify the rows and columns to read. Cross-check important numbers against the original image.
How can I limit the response length?
Use max_tokens to specify the output limit, and describe the required format in your question.
How can I verify the usage of a specific API call?
Read the usage field in the response and verify the usage record for that request in the console.
