At a Glance
| Vendor | Z.ai (Zhipu AI) |
| Family | GLM series |
| Launched | Z.ai released GLM-5.2 on June 13, 2026. It shipped first to paying GLM Coding Plan customers, with open weights following days later on Hugging Face (zai-org/GLM-5.2) under MIT with no regional restrictions. |
| Context window | 1,000,000 tokens, with 128,000 maximum output tokens. |
| Pricing | Z.ai list pricing is approximately $1.40 per million input tokens and $4.40 per million output tokens. The GLM Coding Plan runs at roughly $10 (Lite), $30 (Pro), and $80 (Max) per month, which is where most of its coding-agent usage sits. |
| Access channels | Z.ai API, the GLM Coding Plan, Hugging Face open weights, and self-hosting via transformers, vLLM, SGLang, KTransformers, and Ascend NPU. It drops into Claude Code or Cline with a configuration change, which has driven much of its Western adoption. |
Notable Benchmarks
753B total parameters with approximately 40B active per token. GLM-5.2 scores 62.1 on SWE-bench Pro, ahead of GPT-5.5 at 58.6, and 81.0 on Terminal-Bench 2.1 against Opus 4.8's leading 85.0. It ranks first among open-weight models on the Artificial Analysis Intelligence Index at 51, ahead of Nemotron 3 Ultra (48), MiniMax M3 (44), DeepSeek V4 Pro (44), and Kimi K2.6 (43).
Strengths
Unrestricted MIT licensing with no regional carve-outs, best-in-class open-weight coding scores, 1M context, Ascend NPU support that decouples it from NVIDIA supply, and drop-in compatibility with existing Claude Code workflows at roughly a sixth of the cost.
Limitations
753B parameters still requires serious hardware to self-host despite sparse activation. Terminal-Bench 2.1 trails Opus 4.8. Western enterprise procurement friction around Chinese-origin models applies here as elsewhere, notwithstanding the permissive licence.
Brand-Visibility Implications
GLM-5.2 is the model most likely to be sitting behind a coding agent that a developer configured themselves, because the Claude Code drop-in path and the sixth-of-the-cost pricing make the switch nearly frictionless. For developer-tool brands specifically, that means a growing share of "which library should I use" answers is being generated by GLM rather than by the platform you are monitoring. See Zhipu GLM lineage and Chinese open-source LLM leaderboard.
How Presenc AI Tracks This Model
Presenc AI monitors brand visibility on Z.ai (Zhipu AI)'s GLM series as part of continuous multi-platform AI visibility tracking. We sample Z.ai GLM-5.2 across representative prompt sets daily, compare against competitor performance on the same prompts, and flag material mention-rate changes so brand teams can respond quickly when AI representation shifts.