Zhipu's GLM-5.3-Highspeed Appears in ZCode Gray Release, With Users Reporting About 490 Tokens/s

Back to Devoured - October 13, 2026

DEVOURED

Zhipu's GLM-5.3-Highspeed Appears in ZCode Gray Release, With Users Reporting About 490 Tokens/s

Zhipu AI is testing its new GLM-5.3-Highspeed model via a gray release, with early user benchmarks reporting speeds of approximately 490 tokens per second.

What: The model is currently available to a subset of subscribers within the ZCode platform, focusing on high-throughput inference capabilities.

Why it matters: Rapid token generation is becoming a primary competitive lever for LLM providers as they prioritize latency-sensitive enterprise and real-time agentic applications.

Original article

The model is being rolled out as a gray release to randomly selected subscribers.

Explore all 41 articles in today's digest