Briefs
Briefs
Mar 31
Zhipu AI unveiled GLM-5.1, a coding-focused model that approaches top-tier proprietary performance while pushing a low-cost, open-source strategy for developers and startups.
Zhipu AI has introduced GLM-5.1 and opened the model to existing users of its GLM coding plan.
In coding evaluations run inside the Claude Code framework, GLM-5.1 scored 45.3 compared with 47.9 for Claude Opus 4.6, putting it at roughly 94.6% of that result.
The comparison is notable because the benchmark environment is already optimized for Claude-family workflows, which suggests the gap could narrow further in other coding setups.
Zhipu says the release marks a sharp jump from GLM-5, which scored 35.4 only a month earlier, highlighting a fast pace of model iteration.
Price is a central part of the pitch. The GLM coding plan starts at $3 per month under promotion and is listed at around $10 per month, far below similarly positioned commercial coding models.
Early users cited stronger long-horizon workflow handling, multi-step debugging, and simple setup inside existing developer environments such as Claude Code, where the model can be selected directly in configuration.
By combining near-frontier coding performance with aggressive pricing and an open-source posture, Zhipu is positioning GLM-5.1 as a potential catalyst for a broader shift in the coding-model market.