Chat

GLM-4.5-Air API

Name: GLM-4.5-Air
SKU: glm-4-5-air-awq

GLM-4.5-Air (106B MoE, 12B active) AWQ 4-bit on vLLM. Flagship-class open model, MIT-licensed; fits a single H100 80GB.

$1.50 / 1M tokens · read from the live catalogue

GLM-4.5-Air costs $1.50 / 1M tokens on Open LLMs. Call it by sending glm-4-5-air-awq as the model id to /v1/chat/completions, using any OpenAI-compatible client.

What are GLM-4.5-Air's specifications?

Every value here is read from the catalogue that serves the API, so this page and the service cannot disagree.

Model id: glm-4-5-air-awq
Task: Chat
Price: $1.50 / 1M tokens
Context window: 33K tokens
Modality: text → text
Parameters: 106B (12B active)
Licence: MIT
Endpoint: /v1/chat/completions

How do I call GLM-4.5-Air?

Point an OpenAI client at our base URL and name the model. Nothing else about your request changes.

GLM-4.5-Air — first requestbash

curl https://api.openllms.co/v1/chat/completions \
  -H "Authorization: Bearer $OPENLLMS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-4-5-air-awq",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Before your first call

An inference call needs an API key and a positive wallet balance. The first call to an idle model can answer 503 with a Retry-After header while it starts — wait, then retry. Full reference: the API docs.

How is GLM-4.5-Air licensed?

GLM-4.5-Air is served under the MIT licence. The upstream licence passes through to you: what you may do with the output is governed by it, not by us. The obligations we pass through are listed on model licences, and the model card itself is at its source.

What else can I call?

See every model in the catalogue with its price, context window and licence.