BLAST RADIUS

Open-weight & owned stack · Quantization efficiency

Zhipu Reports Fast GLM Inference On Chinese Chips

Sep 18, 2026. Covered by 2 outlets.

Zhipu said GLM-5.3-FlashX reached nearly 200 tokens per second across roughly 100,000 domestic AI accelerators.

Read the original at finance.biggo.comOpens the publisher's site in a new tab

Also covering this

Zhipu Opens GLM-5.3-FlashX Near 200 Tokens/s on ~100k Domestic Accelerators - Pandailypandaily.com, Sep 18

More in Open-weight & owned stack

Xiaomi Publishes MiMo-V2.6 Reinforcement Learning DashboardSep 17Gemma Models Surpass One Billion DownloadsSep 16llama.cpp Fixes Remote-Code-Execution VulnerabilitySep 16Alibaba Releases Occamy Open-Weight AI AgentSep 15Apple Enables Local Foundation Models On MacOSSep 15