GLM-5.2-Fast-Preview is a high-speed variant of Zhipu AI's GLM-5.2.
It is described as an inference-optimized preview model with capabilities on par with the standard version.
| Developer | Zhipu AI |
|---|---|
| Model ID | zhipu/glm-5-2-fast-preview |
| Release Date | July 9, 2026 |
| Context Window | 1,000,000 tokens (~1M) |
| Max Output Tokens | 131,072 tokens |
| Modalities | Text → Text |
| Capabilities | parallel tool calls, reasoning, streaming, structured (JSON schema) output, tool/function calling |
High-speed variant
The model is positioned as the fast-inference version of GLM-5.2.
Latency-focused
It is described as suitable for latency-sensitive scenarios.
Output throughput
The model is said to deliver 1.5–2× the output TPS of the standard version.
Service provider
The inference service provider for glm-5.2-fast-preview is listed as Alibaba Cloud Model Studio.
Model identity
Search results consistently identify it by the exact model id glm-5.2-fast-preview.
Real-time chat
It is described as fitting real-time dialogue and chat workloads.
Multi-turn agents
It is described as suitable for multi-turn agent interactions.
Streaming code generation
It is described as a fit for streaming code generation.
Speed-sensitive inference
It is positioned for scenarios where output speed is critical.
Sources: https://www.qwencloud.com/models/glm-5.2-fast-preview · https://cloudprice.net/models/zhipu-glm-5-2-fast-preview · https://aihubmix.com/compare/glm-5.2/glm-5.2-fast-preview
Measured on the AI/ML API gateway across live traffic. Figures are medians unless stated. Last updated 2026-10-12.
Real-time chat is a documented fit for this model because it is inference-optimized for latency-sensitive use cases.
Multi-turn agents are also explicitly listed as a suitable use case, including agent-style interactions that benefit from fast output.
Streaming code generation is another documented use case for this model, reflecting its optimization for low-latency, incremental output.
The model is also described as suitable for real-time dialogue and multi-turn Agent calls in another source covering the same exact model.
import requests r = requests.post( "https://api.aimlapi.com/v1/chat/completions", headers={"Authorization": "Bearer " + AIMLAPI_KEY}, json={ "model": "zhipu/glm-5-2-fast-preview", "messages": [ { "role": "user", "content": "Hello!" } ] }, ) print(r.json())
const r = await fetch("https://api.aimlapi.com/v1/chat/completions", { method: "POST", headers: { Authorization: `Bearer ${process.env.AIMLAPI_KEY}`, "Content-Type": "application/json", }, body: JSON.stringify({ "model": "zhipu/glm-5-2-fast-preview", "messages": [ { "role": "user", "content": "Hello!" } ] }), }); console.log(await r.json());
curl -X POST https://api.aimlapi.com/v1/chat/completions \ -H "Authorization: Bearer $AIMLAPI_KEY" \ -H "Content-Type: application/json" \ -d '{"model":"zhipu/glm-5-2-fast-preview","messages":[{"role":"user","content":"Hello!"}]}'
OpenAI-compatible — swap the base URL and it works with your existing SDK.
| Type | Price |
|---|---|
| Input | |
| Output | |
| Cached input |
glm-5.2-fast-preview is documented with a cache-hit price alongside input and output prices, at CNY 4 per million tokens in one provider listing.
| Model | Input | Output | Context | Best for |
|---|---|---|---|---|
GLM 5.2 Fast Preview This page | Agentic workflows and structured output |
GLM 5.2 Fast Preview has a 1,000,000 tokens context window and can return up to 131,072 tokens.
GLM 5.2 Fast Preview takes text as input and returns text.
Use zhipu/glm-5-2-fast-preview as the model id. Requests go to https://api.aimlapi.com/v1/chat/completions.
GLM 5.2 Fast Preview became available on July 10, 2026.
GLM 5.2 Fast Preview is priced at input $4.55 / 1M tokens, output $14.3 / 1M tokens, cached input $0.91 / 1M tokens.
Yes, GLM 5.2 Fast Preview can stream responses as they are generated.
GLM 5.2 Fast Preview was built by Zhipu AI.
It is available at the endpoint https://api.aimlapi.com/v1/chat/completions using the model id zhipu/glm-5-2-fast-preview.
It supports reasoning, streaming, structured output, tools, and parallel tool calls.
