Zhipu AI Open-Sources GLM-5.3-Flash Native Multimodal Model, Priced at 1/20 of GLM-5.3
Summary
Zhipu AI tonight released and open-sourced GLM-5.3-Flash, the first native multimodal model in the GLM-5 series, with 320B total parameters and only 18B active parameters. The model surpasses GLM-5.2 on multiple benchmarks, matches Claude Opus 4.8, yet is priced at 1/10 of GLM-5.3, with a limited-time discount of 1/20. Its architecture is designed for extremely low cost, using a hybrid sparse attention and linear attention architecture, and all compute is powered by domestic chips.
Bubble analysis
This news is not directly about the investment bubble, but it reflects rapid progress in cost efficiency of Chinese AI models, which could affect global AI investment return expectations. If models can achieve frontier performance at lower cost, it may reduce reliance on massive capex, offering some counter-evidence to the bubble thesis.
Related signals
linked by shared tags
- WeChat Open-Sources Universal Multimodal Embedding Model WeMM-Embedding, Deployed at Scale in WeChat Recommendation and Search Systems#open-source#multimodal
- Zhipu GLM-5.3-Flash Launches with SenseTime's Domestic Computing Power Support#zhipu#glm
- GLM-5.3-Flash will likely handle 45% of your AI workloads#open-source#glm
- DeepSeek, Qwen, and Zhipu Take Turns on Stage; PC Makers Finally Get Their Ammunition#zhipu
- Alibaba Qwen Open-Sources Qwen-Drive-1.0-4B: First Vision-Language Foundation Model for Autonomous Driving#open-source