Zhipu AI's GLM-5.2 rivals Claude on extended coding tasks

Zhipu AI releases GLM-5.2, an open-source model under MIT license with stable 1-million-token context window

Published 8sem1 sourceNotable

≈ 21sSpeed

The fact

On FrontierSWE benchmark for extended coding tasks, GLM-5.2 trails Anthropic's Claude Opus 4.8 by just 1 percentage point

The model still lags significantly behind closed-source rivals on reasoning capabilities

Click the link to read an article on the topic:

Explore this topic

Auto-synthesis from 1 media source · identified on June 17, 2026