Cheap vision with tool calling for apps that read screenshots and documents; the Flash version costs nothing in the API.
Kenning's take
What it does
GLM-4.6V is a multimodal model from Z.ai that takes images, video and text and answers in text. Its main addition is native function calling, so it can use tools while working with visual input. It is available through the Z.ai API and as open weights, alongside a smaller GLM-4.6V-Flash version.
- Image, video and text input
- Native tool calling for visual tasks
- 128K-token context
- Smaller GLM-4.6V-Flash variant
- MIT-licensed open weights
Who it suits. Developers building apps that need to understand screenshots, documents, images or video.
Key facts
- Category
- AI models
- Made by
- Z.ai (Zhipu AI)
- Based in
- China
- Released
- 8 December 2025
- How to get it
- Z.ai API (model glm-4.6v) or self-hosted from open weights on Hugging Face; GLM-4.6V-Flash is free on the API
- Works on
- API, open weights (Hugging Face)
- In the UK
- Yes
- Free plan
- GLM-4.6V-Flash is free on the Z.ai API; weights free to download
- API price
- $0.30 per million input tokens, $0.90 per million output tokens
Reviews and write-ups
- TestingCatalog 2025-12-08
FAQ
When was GLM-4.6V released?
Z.ai announced GLM-4.6V on 8 December 2025.
Is GLM-4.6V open source?
Yes, the weights are on Hugging Face under the MIT licence.
How much does the GLM-4.6V API cost?
$0.30 per million input tokens and $0.90 per million output tokens; the Flash version is free.
Facts from the maker's own pages and the sources listed. Checked 2026-10-11.
Similar tools
ACE-Step 1.5
Open-source music generation model that writes full songs with lyrics and runs on ordinary consumer GPUs.
AlphaGenome Atlas
Google DeepMind database of AI-predicted effects for every possible single-letter change in human DNA.
Amazon Nova 2
Amazon's second family of Nova AI models, covering reasoning, speech conversation and combined text-and-image generation.
Cisco Antares
Cisco's small open-weight models that find where known security vulnerabilities sit in a codebase, running locally.