GLM-5.1 (Z.ai)

GLM-5.1 is a text model from Z.ai, formerly Zhipu AI, positioned for software-development work within the GLM Coding Plan. The available launch report says it was made available to all plan users, establishing a clear coding-oriented use case without attributing to it an architecture, training method, parameter count, or context design that the supplied evidence does not establish.

In one reported coding evaluation conducted with Claude Code as the testing tool, GLM-5.1 scored 45.3, compared with 35.4 for its predecessor GLM-5 and 47.9 for Claude Opus 4.6. The report characterizes this as a 28% improvement over GLM-5 and 94.6% of the Opus score, suggesting strong practical potential for coding tasks while treating the single-source benchmark as reported rather than independently reproduced evidence.