# Zhipu Launches GLM-5.1 High-Speed API, GLM-5.1-highspeed

- Category: Flash (快讯)
- Published: 2026-05-22T12:00:00.000Z
- Source date: 2026-05-22
- HTML: https://www.88z.ai/en/news/flash-20260522-8d4410cbaefddb1f
- Markdown: https://www.88z.ai/en/news/flash-20260522-8d4410cbaefddb1f.md
- Other language: https://www.88z.ai/news/flash-20260522-8d4410cbaefddb1f.md

Zhipu Open Platform has released the GLM-5.1 High-Speed API, achieving an output speed of 400 tokens/s and setting a new global benchmark for large model API speed. Jointly launched by Zhipu and the TileRT team, it combines flagship capabilities with ultra-low latency. It is suitable for high-demand scenarios like AI coding and real-time interaction, and is now available to select enterprise clients.

## At a glance

- GLM-5.1 High-Speed API reaches 400 tokens/s, setting a new global record
- Jointly launched by Zhipu and TileRT, balancing flagship capabilities with low latency for select enterprises

## The update

Zhipu Open Platform has officially launched the GLM-5.1 High-Speed API (GLM-5.1-highspeed), with a model output speed reaching 400 tokens/s, thereby setting a new global record for large model API speed. This breakthrough was achieved jointly by Zhipu and the TileRT team, marking the first time a domestic large model has combined flagship-level intelligence with ultra-low latency. This version is designed for applications highly sensitive to latency, such as AI programming assistance, real-time human-computer interaction, and real-time voice processing. By optimizing the underlying architecture, GLM-5.1-highspeed significantly reduces response time while maintaining high-performance output. Currently, the service is open to select enterprise clients, aiming to meet the strict dual demands for speed and quality in high-end business operations.

## Sources

- [智谱开放平台](https://mp.weixin.qq.com/s/TAVXB_Xuys7_R-SG4oFhOA) (2026-05-22)

This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.