GLM5.2 model: openweight coding performance data
GLM-5.2 improved internal task success rates from 21/70 to 48/70 over its predecessor, signaling a shift in open-weight viability.
Marcus Chen
13 min read