發布於 2026-08-16 · 赵予安
直接回答
智谱 GLM 5.2 CN 標價為每百萬 token 0.171 / 0.596。上下文窗口和該精確 ID 的 CacheSafety 仍未公布。 這篇指南面向正在比較模型品質、成本、路由策略和正式環境上線風險的台灣產品與平台團隊。
What shipped
Z Ai has a new model entry in the NextModel catalog: GLM 5.2 CN. The public model slug is z-ai--glm-5.2-cn. The catalog model id is z-ai/glm-5.2-cn. The provider is Z Ai. The display name is GLM 5.2 CN. Streaming is a listed capability. The latency band is 1000-3000ms. The catalog does not publish a context window for this model.
Pricing
The input price is $0.171 per 1M tokens. The output price is $0.596 per 1M tokens. Related catalog entries show GLM 5.1 CN at $0.120 per 1M tokens input and $0.479 per 1M tokens output. GLM 5 CN is listed at $0.084 per 1M tokens input and $0.373 per 1M tokens output.
Who it is for
The catalog lists use cases as quality. That makes this model a quality-oriented option for workloads selected through the NextModel gateway. Streaming is available, so the model has a place in applications that need incremental output.
What is unverified
The release date is unpublished in this catalog snapshot. The catalog value for context window is 0, so it is unpublished. No first-party CacheSafety Bench row exists for this exact model. Safe Hit, Bad Hit, and trap rates are not included in these notes. Independent third-party quality scores are not in these notes.