# `Agentic.LLM.Provider.Zai`

z.ai (formerly Zhipu) Provider — OpenAI-compatible.

z.ai serves the GLM family. The wire format is OpenAI-compatible
(`Authorization: Bearer <KEY>`), so this provider reuses the OpenAI
Chat Completions transport with a different base URL.

## Endpoints

  * Global (USD billing): `https://api.z.ai/api/paas/v4/`
  * China (CNY billing): `https://open.bigmodel.cn/api/paas/v4/`
  * Coding-Plan keys: `https://api.z.ai/api/coding/paas/v4/`

We default to the global endpoint. Hosts that need the Coding-Plan
variant should override `default_base_url` via the user-config
override path.

z.ai exposes **no `/models` endpoint** and **no balance/quota endpoint**;
the model list is static and per-account quotas are dashboard-only at
https://z.ai/manage-apikey/billing.

Model pricing per 1M tokens (USD):

    glm-4.5         $0.60 / $2.20    131k ctx
    glm-4.5-air     $0.13 / $0.85    131k
    glm-4.5-flash   $0.00 / $0.00     128k  (free tier; rate-limited)
    glm-4.6         $0.39 / $1.74    205k
    glm-4.7         $0.38 / $1.74    203k
    glm-4.7-flash   $0.06 / $0.40    203k

---

*Consult [api-reference.md](api-reference.md) for complete listing*