GLM-5.3-FlashX · 10M tokens

The same capability profile as Flash, tuned for inference speed at around 200 output tokens per second, for interfaces where perceived latency matters.

What you get

• 10M tokens for every unit you buy. Order several units to stack quota.
• A personal access key for glm-5.3-flashx, issued automatically once your order is completed.
• No shipping: delivery is digital, under My Account > Order History and by email.

Capabilities

• Fastest
• 200 tok/s
• Multimodal
• 1M context

Best for

Latency-sensitive agents and interactive multimodal tools.

Quick start

from zai import ZaiClient

client = ZaiClient(api_key="YOUR_ACCESS_KEY", base_url="YOUR_BASE_URL")
response = client.chat.completions.create(
    model="glm-5.3-flashx",
    messages=[{"role": "user", "content": "Hello"}],
    reasoning_effort="max",
)

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