Z.ai debuts open source GLM-4.6V, a native tool-calling vision model for multimodal reasoning
Chinese AI startup Z.ai has released its GLM-4.6V series, a new generation of open-source vision-language models optimized for multimodal reasoning, frontend automation, and high-efficiency deployment.
Why it matters
The GLM-4.6V series introduces native multimodal tool calling capabilities, enabling more efficient and powerful multimodal reasoning and automation.
Key Points
- 1Z.ai released two models: GLM-4.6V (106B) and GLM-4.6V-Flash (9B)
- 2The models support native function calling, enabling direct use of tools like search, cropping, or chart recognition with visual inputs
- 3The models have a 128,000 token context length and state-of-the-art results across 20+ benchmarks
- 4The models are available under a permissive MIT license, making them suitable for enterprise adoption
- 5The models use a Vision Transformer encoder and support arbitrary image resolutions and temporal sequences
Details
The GLM-4.6V series from Chinese AI startup Z.ai introduces native function calling in a vision-language model, allowing direct use of tools like search, cropping, or chart recognition with visual inputs. The series includes two models - a larger 106-billion parameter GLM-4.6V for cloud-scale inference, and a smaller 9-billion parameter GLM-4.6V-Flash for low-latency, local applications. The models have a 128,000 token context length and state-of-the-art results across over 20 benchmarks, positioning them as a competitive alternative to closed and open-source vision-language models. The models use a Vision Transformer encoder and support arbitrary image resolutions and temporal sequences. They are available under a permissive MIT license, making them suitable for enterprise adoption, including scenarios requiring full control over infrastructure, compliance with internal governance, or air-gapped environments.
No comments yet
Be the first to comment