Zhipu Open-Sources GLM-5.3-Flash Multimodal Model

Zhipu AI released and open-sourced GLM-5.3-Flash, the first native multimodal model in the GLM-5 series. It has 320 billion total parameters but only 18 billion active. Before launch, it was tested anonymously as Ox-Alpha, quickly becoming the most popular model on OpenCode and OpenRouter, setting new usage records, with all computing powered by domestic chips.

GLM-5.3-Flash outperforms GLM-5.2, scoring 57 on the AA Intelligence Index, matching Anthropic's Claude Opus 4.8, with similar coding performance. Its price is one-tenth of GLM-5.3, one-twentieth during discounts, and one-fortieth of Opus 4.8.

The model uses a hybrid sparse and linear attention architecture, a first for open-source frontier models, greatly reducing long-context serving costs. Vision capabilities are natively integrated into the coding loop, allowing it to observe interfaces and improve. It also extends to Office and financial research, autonomously completing workflows from analysis to document delivery.

GLM-5.3-Flash is now globally open-sourced, available on ZCode and included in the GLM Coding Plan, with API access open.