GLM-5.1 Beat GPT-5.4 on SWE-Bench Pro — Did China Just Win the Coding War?

Añadido: 2026-05-21

GLM-5.1, a Chinese-developed large language model, has demonstrated superior performance on SWE-Bench Pro, a comprehensive coding benchmark that evaluates models' ability to fix real software bugs, outperforming leading US models including GPT-5.4 and Claude Opus 4.6. This achievement highlights the competitive landscape of AI coding models and demonstrates that Chinese AI development has reached parity or leadership in advanced coding capabilities.