Search
Zhipu Unveils New Generation of Open Source Models, Boasting Speeds of Up to 200 Tokens/Second for Turbo Variant
2025-04-15 /
Read about 0 minute
Author：小编
Zhipu has introduced a pioneering generation of open source models, the GLM-4-32B-0414 series, encompassing base, inference, and deliberation model weights. Licensed under the MIT License, these models are readily accessible and experiencable via the "z.ai" platform. Notably, the inference model GLM-z1-Air/AirX-0414 distinguishes itself with its blistering speed of up to 200 tokens per second, positioning it as the swiftest commercial model currently available in China. Furthermore, its cost is a mere fraction of DeepSeek-R1's, specifically one-thirtieth.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
