Zhipu Releases GLM-5.3-Flash: GPT-4o Mini Rival with 50% Faster Inference

Philpax · hn · 2026-08-26

Zhipu AI released the GLM-5.3-Flash model, focusing on low cost and high performance. The model significantly reduces inference costs while maintaining high capabilities, with a 50% speed improvement over the previous generation. It claims performance comparable to GPT-4o mini, supports 128K context, and is priced at 0.1 RMB per million tokens.

Original post →

More from Models

Models channel →