KeyoAPI
GLM 5.2 API — Call Zhipu GLM on KeyoAPI (OpenAI-Compatible)
Developers searching for a GLM 5.2 API usually want two things at once: a current Zhipu-class model and an OpenAI-shaped client they already trust.
Overview
Developers searching for a GLM 5.2 API usually want two things at once: a current Zhipu-class model and an OpenAI-shaped client they already trust. KeyoAPI exposes glm-5.2 as a paid twin and glm-5.2-free as a permanent $0 fair-use twin on the same gateway. That combination matters for teams evaluating glm api options after a new release window: you can prototype tonight on free, then flip one string to the token-metered ID when traffic hardens. GLM 5.2 sits in the general-purpose chat and reasoning lane — product copilots, bilingual support replies, structured extraction, and agent tool loops that already speak /v1/chat/completions. On KeyoAPI you do not juggle a separate Zhipu console for every environment. One prepaid balance covers glm-5.2 next to DeepSeek, Kimi, GPT-class, Whisper, and vision models. That is the practical meaning of an AI API relay for overseas developers who already standardized on OpenAI SDKs in Python, Node, or Cursor-compatible clients. Indicative Keyo sell rates for glm-5.2 are about $3.50 per 1M input tokens and $11.00 per 1M output tokens. Treat catalog numbers as planning anchors and confirm live rates on the interactive /pricing/glm-5.2 page before you sign a volume quote. Unlike trial coupons that expire into a paid plan, glm-5.2-free keeps ModelPrice at $0; the trade-off is fair-use rate limiting and lower priority on the Keyo Free channel. If your SLA needs stable concurrency, keep paid glm-5.2 in production and reserve free for CI smoke tests, demos, and eval harnesses. Why this page targets glm 5.2 api and the broader glm api cluster: model-release windows create short-lived search spikes when engineers compare vendors, paste curl snippets, and look for a key they can mint in minutes. Hugging Face threads and official docs will own brand-heavy discussion. Keyo’s job is different — show the exact model ID, the dual free/paid path, a honest fair-use caveat, and how pricing stacks against staying on a single-vendor bill. That is enough for long-tail conversion even when you are not chasing position one on the naked model name. Common shipping patterns: (1) bilingual SaaS support where Chinese and English tickets share one agent stack; (2) document-to-JSON pipelines that pair OCR on Keyo with glm-5.2 for field filling; (3) overnight batch jobs that start on free IDs during dry runs, then switch to paid for the full corpus. Because the base URL stays https://www.keyoapi.xyz/v1, environment promotion is a config change, not a rewrite. Integration is intentionally boring. Export OPENAI_BASE_URL=https://www.keyoapi.xyz/v1 and OPENAI_API_KEY, then call chat completions with model=glm-5.2. Streaming, tools, and JSON modes follow the same shapes your OpenAI client already expects. When you need a cheaper high-volume lane, compare gpt-5.6-luna on the same key; when you want another Chinese-capable family, try deepseek-v4-flash or kimi-k3. See also /free-models for the full $0 catalog and /deepseek-api-pricing if you are shopping across families before you lock a default model. Bottom line: KeyoAPI’s GLM 5.2 API path is built for developers who want glm api access without a second billing island — permanent free for tryout, transparent token rates for scale, and one relay for the rest of the multimodal stack.
Free tier available
Same model family on KeyoAPI: use glm-5.2-free for permanent $0 fair-use calls (Keyo Free channel), or glm-5.2 for token-metered production traffic.
Rules: free = *-free suffix; paid = bare ID. Fair-use rate/concurrency limits apply on free — see /free-models.
Free curl
curl https://www.keyoapi.xyz/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"glm-5.2-free","messages":[{"role":"user","content":"Hello"}]}'Paid curl
curl https://www.keyoapi.xyz/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"glm-5.2","messages":[{"role":"user","content":"Hello"}]}'Quick start
Base URL: https://www.keyoapi.xyz/v1
Endpoint: POST /v1/chat/completions
Model: model=glm-5.2
curl https://www.keyoapi.xyz/v1/chat/completions \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"glm-5.2","messages":[{"role":"user","content":"Summarize GLM 5.2 API pricing on KeyoAPI"}]}'Auth uses the same Bearer API key as chat. Full notes: Keyo docs.
FAQ
What is the GLM 5.2 API model ID on KeyoAPI?
Use model=glm-5.2 for the paid, token-metered twin. For the permanent $0 fair-use tier, use model=glm-5.2-free. Both speak OpenAI-compatible chat completions at https://www.keyoapi.xyz/v1.
How does KeyoAPI GLM API pricing compare to official Zhipu?
Keyo lists glm-5.2 around $3.50 input / $11.00 output per 1M tokens (indicative). Always confirm live sell rates on /pricing/glm-5.2. The free ID is fixed at $0 with fair-use limits.
Can I migrate an existing OpenAI SDK client to glm-5.2?
Yes. Set baseURL to https://www.keyoapi.xyz/v1, keep Authorization: Bearer YOUR_KEY, and change the model string to glm-5.2 or glm-5.2-free. No custom Gemini/Anthropic SDK required for chat.
Is glm-5.2-free really permanent?
Yes — ModelPrice is $0 for the free ID. Fair-use rate and concurrency limits apply on the Keyo Free channel. Production QPS should use the paid bare ID.