Prime mode

Prime mode provides higher TPS for scenarios sensitive to output speed.

Usage

Prime mode provides the following key features:

• High-speed output: TPS is increased to 1.5 to 2 times that of the standard API. This is suitable for AI coding assistants, multi-step Agent reasoning, real-time conversations, and other latency-sensitive scenarios.
• Token-based billing: The billing logic is the same as the standard API, with charges based on input and output tokens.
• Special rate limiting: When the call volume reaches the rate limit, if the platform still has spare resources, rate limiting is not triggered, so the actual available TPS is no lower than the rate limit.

To enable Prime mode, set the model parameter to the model ID of Supported models. No additional parameters are required. The endpoint is in the format https://{workspace_id}.cn-beijing.maas.aliyuncs.com/compatible-mode/v1, where {workspace_id} can be found on the Business Space Management page after you switch to the corresponding region.

Basic call example:

The model name for the prime mode of glm 5.2 is still glm-5.2-fast-preview.

curl -X POST https://{workspace_id}.cn-beijing.maas.aliyuncs.com/compatible-mode/v1/chat/completions \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.2-fast-preview",
    "messages": [{"role": "user", "content": "Who are you"}],
    "stream": false
}'

Supported models

Examples

The glm-5.2 model returns a reasoning_content field by default for reasoning. During streaming, reasoning content and response content are delivered through delta.reasoning_content and delta.content respectively. Streaming call example:

from openai import OpenAI
import os

client = OpenAI(
    api_key=os.environ.get("API_KEY"),
    base_url=os.environ.get("BASE_URL"),
)

completion = client.chat.completions.create(
    model="glm-5.2-fast-preview",
    messages=[{"role": "user", "content": "Who are you"}],
    stream=True,
)

for chunk in completion:
    if not chunk.choices:
        continue
    delta = chunk.choices[0].delta
    if hasattr(delta, "reasoning_content") and delta.reasoning_content:
        print(delta.reasoning_content, end="", flush=True)
    if hasattr(delta, "content") and delta.content:
        print(delta.content, end="", flush=True)

Response example:

{
  "id": "chatcmpl-xxx",
  "object": "chat.completion",
  "model": "glm-5.2-fast-preview",
  "choices": [{
    "index": 0,
    "message": {
      "role": "assistant",
      "reasoning_content": "...",
      "content": "..."
    },
    "finish_reason": "stop"
  }],
  "usage": {
    "prompt_tokens": 14,
    "completion_tokens": 137,
    "total_tokens": 151,
    "prompt_tokens_details": { "cached_tokens": 0 },
    "completion_tokens_details": { "reasoning_tokens": 127 }
  }
}

Billing

For the billing rules and pricing of Prime mode, see Model inference pricing.

Error codes

For the error codes that may occur when you call Prime mode and the corresponding solutions, see Error codes.