GLM 5.3 API: Endpoints, Parameters & Python Examples

Quick answer: The GLM 5.3 API (model id glm-5.3) is rolling out at $1.40 per 1M input / $4.40 per 1M output tokens — the same rate as GLM 5.2. It's OpenAI- and Anthropic-compatible, requires thinking always enabled (reasoning_effort: low/high/max), and supports 1M-token context, 128K output, function calling with streaming parameters, and structured output.

TL;DR

Field	GLM 5.3 API
Model ID	`glm-5.3`
Price	$1.40 in / $4.40 out per 1M ($0.26 cached)
Context / Output	1M tokens / 128K
Thinking	Always on — `reasoning_effort` low/high/max
Protocols	OpenAI Chat Completion / OpenAI Response / Anthropic Message
Status	Rolling out ("coming soon" per official docs)

Endpoints

Z.AI publishes three compatible endpoints (all already live in docs):

Protocol	Base URL
OpenAI Chat Completion	`https://api.z.ai/api/coding/paas/v4`
OpenAI Response	`https://api.z.ai/api/v1`
Anthropic Message	`https://api.z.ai/api/anthropic`

The OpenAI Chat Completion endpoint is the most widely used — and the one that works for previous Coding Plan subscribers (including expired ones) during rollout.

Required Parameters: Thinking Is Always On

The one breaking change vs GLM 5.2: you can no longer disable thinking.

{
  "model": "glm-5.3",
  "thinking": { "type": "enabled" },
  "reasoning_effort": "max"
}

Parameter	Values	Default	Notes
`thinking.type`	`enabled`	`enabled`	`disabled` is rejected
`reasoning_effort`	`low`, `high`, `max`	`max`	`max` recommended for coding

Migration: set enabled + low before switching the model ID, then tune up.

Python Example (OpenAI-Compatible)

from openai import OpenAI

client = OpenAI(
    base_url="https://api.z.ai/api/coding/paas/v4",
    api_key="YOUR_ZAI_API_KEY",
)

resp = client.chat.completions.create(
    model="glm-5.3",
    messages=[
        {"role": "user", "content": "Write a Python function that checks if a number is prime, with a docstring."}
    ],
    extra_body={
        "thinking": {"type": "enabled"},
        "reasoning_effort": "max",
    },
)

print(resp.choices[0].message.content)

Pricing

Model	Input/1M	Cached/1M	Output/1M
GLM 5.3	$1.40	$0.26	$4.40
GLM 5.2	$1.40	$0.26	$4.40
Kimi K3 (ref)	$3.00	–	$15.00

At $0.68 per Artificial Analysis Intelligence Index task, GLM 5.3 is the cheapest top-tier model per unit of measured intelligence — though it's also the most verbose of the group (~170M output tokens across AA's eval suite vs 72M median).