Back to Models
GLM 4.5 Air
zhipuzhipu/glm-4.5-air
Cost-optimised GLM chat model with siliconflow + zhipu fallback.
glm-4.5-air is the lightweight member of the GLM-4.5 family β strong reasoning and tool use at a fraction of the flagship's price. As of 2026-05, it bills at $0.15 per million input tokens and $1.20 per million output tokens on TheRouter.
TheRouter primarily routes glm-4.5-air through siliconflow-intl (priority 0) and falls back transparently to BigModel direct (zhipu-cn, priority 1) if siliconflow returns a 5xx or times out. You make a single OpenAI-compatible call; we handle the failover.
Lightweight reasoning
Reasoning, function calling, JSON mode, and prompt caching at small-model latency.
Dual-provider routing
Siliconflow primary, BigModel direct fallback β failover is silent and automatic.
Strong Chinese
Native Chinese-language performance from a model trained primarily on Chinese corpora.
OpenAI-compatible
Same `/v1/chat/completions` endpoint, same SDKs β just change the model ID.
When to use
High-volume Chinese-leaning chat workloads where cost per token matters: customer support, content rewriting, summarisation, agent orchestration.
When not to use
Pick a flagship (glm-4.7, claude-sonnet, gpt-5) for hard reasoning, complex code, or English-dominant high-stakes tasks.
Pricing: $0.15 input / $1.20 output per MTok. ~26% margin over CN cost at the locked rate.
Read the glm-4.5-air tutorial βContext Length
131K
Max Output
66K
Input Priceper 1M tokens
$0.180/ 1M tokens
Output Priceper 1M tokens
$1.44/ 1M tokens
Modalities
textβtext
Pricing Breakdown
| Type | Rate |
|---|---|
| Input | $0.180 / 1M tokens |
| Output | $1.44 / 1M tokens |
Supported Parameters
temperaturemax_tokenstop_ptoolstool_choicestop
API Usage Examples
Use the global api.therouter.ai endpoint shown below for new integrations; the legacy China accelerated endpoint is retired.
cURL
curl https://api.therouter.ai/v1/chat/completions -H "Content-Type: application/json" -H "Authorization: Bearer $THE_ROUTER_API_KEY" -d '{
"model": "zhipu/glm-4.5-air",
"messages": [
{"role": "user", "content": "Summarize the key points from this input."}
]
}'