Zhipu AI's new - generation model GLM - 4 - 32B - 0414 series

GLM-4-32B-0414 Series of the New Generation of Models Open-Sourced by Zhizhi AI

Zhizhi AI has released the GLM-4-32B-0414 series of the new generation of open-source models, which includes base, inference, and contemplation model weights and follows the MIT License. Currently, all models in the series can be accessed and experienced via "z.ai". According to the official introduction of Zhizhi AI, the inference models GLM-Z1-Air/AirX-0414 can achieve a maximum inference speed of 200 Tokens per second, which is the fastest commercial model in China so far. Moreover, the price of GLM-Z1-Air-0414 is only 1/30 of that of DeepSeek-R1. In addition, on April 14, 2025, Zhizhi AI, a unicorn in the large language model field, filed for guidance with the Beijing Securities Regulatory Bureau, and the guidance institution is China International Capital Corporation Limited. This means that Zhizhi AI has become the first among the "Six Dragons of Large Language Models" to officially launch the IPO process.

—— China Star Market Daily, Securities Times

via Windvane Reference Express - Telegram Channel