FAST GLM-4.5128K CONTEXTPROMPT CACHINGTEXT + CODEUSD PRICING
GLM-4.5 AirX API Pricing
GLM-4.5 AirX is the fast mid-tier GLM-4.5 SKU between Air and X. The live Z.AI pricing table lists $1.1/M input and $4.5/M output, with cached input at $0.22/M. Pulled directly from docs.z.ai daily.
Input - per 1M tokens
$1.10/M
Source Z.AI flat
Output - per 1M tokens
$4.50/M
Context 128K flat
Cached input - per 1M tokens
$0.22/M
Storage limited-time free -80%
Effective - agentic blend
$0.71/M
92/8 split - 82% cache
§ 01 / TERMINAL
Run the numbers.
Live calculator pre-loaded with current GLM-4.5 AirX rates. Tweak spend, output mix, or cache assumptions and share the URL to share the calculation.
$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
§ 02 / SCENARIOS
Real-world presets.
CODING AGENT
Small repo implementation
$0.038/task
CODE REVIEW
Pull request review
$0.024/review
RAG
Knowledge base answer
$0.014/query
CHATBOT
Support assistant
$0.005/turn
§ 03 / TAPE
Price history.
Input · $1.1/M
Output · $4.5/M
Cached · $0.22/M
JUL 28 Launch-wave list price stored at $1.1/M input and $4.5/M outputMAY 19 Live verification kept $1.1/M input and $4.5/M output
§ 04 / TOKENIZER
Paste text. See tokens. See cost.
Estimate · zhipu-tokenizer-estimate · ≈3.85 chars/token Auto-counts as you type
This is a chars-per-token approximation, not a real tokenizer. Actual tokens vary by language, code density, and tool-call overhead — counts are typically ±10–20% off for English prose, more for code or non-Latin scripts. For exact billing, use the vendor's official tokenizer.
Characters —
Words —
Tokens (estimated) —
Cost as input · uncached —
Cost as output · uncached —
Cost as cached input —
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| GLM-4.5 AirX Current | $1.10 cache $0.22 | $4.50 | $0.71 agentic 92/8 | 128K | Fast GLM-4.5 production |
| GLM-4.5 X | $2.20 cache $0.45 | $8.90 | $1.42 pricier | 128K | Harder GLM-4.5 reasoning |
| GLM-4.5 Air | $0.20 cache $0.03 | $1.10 | $0.14 cheaper | 128K | Budget GLM-4.5 production |
| GLM-4.7 | $0.60 cache $0.11 | $2.20 | $0.36 cheaper | 200K | Mid-tier GLM agents |
| GLM-5 | $1.00 cache $0.20 | $3.20 | $0.57 cheaper | 200K | Balanced GLM flagship |
| GLM-4.7 FlashX | $0.07 cache $0.01 | $0.40 | $0.05 cheaper | 200K | Ultra-cheap GLM traffic |
| DeepSeek V4 Flash | $0.14 cache $0.00 | $0.28 | $0.05 cheaper | 1M | Budget reasoning and coding |
| Qwen 3.5 Flash | $0.10 | $0.40 | $0.12 cheaper | 1M | Bulk Qwen long-context work |
Frequently asked.
Practical pricing questions, separated from calculator assumptions and regional tiers.
Q · 01 What is GLM-4.5 AirX priced at? +
GLM-4.5 AirX is listed at
$1.1/M input and $4.5/M output on the live Z.AI pricing table. Cached input is listed at $0.22/M. This page stores USD per-million-token pricing.Q · 02 How is the effective price calculated? +
AI//COST uses the same 92/8 agentic blend everywhere. With an 82% cache hit rate, cached input uses the vendor cache-read row. GLM-4.5 AirX's effective blended cost is
$0.71/M.Q · 03 Is prompt caching priced separately? +
Yes. The vendor table lists cached input at
$0.22/M versus $1.1/M fresh input. Cached input storage is listed as limited-time free on the Z.AI page.Q · 04 Are regional prices different? +
Z.AI publishes the official developer pricing page in USD. Chinese BigModel pages may surface overlapping model catalogs, but the quote tiles use the baseline row from the Z.AI developer pricing page, not a reseller or proxy price.
Q · 05 Is there a batch discount? +
No separate batch discount row is listed on the Z.AI pricing page for GLM-4.5 AirX. The quote tiles show real-time list pricing; batch economics should be treated as a separate calculator variant only when the vendor documents it.
Q · 06 How accurate is the tokenizer estimate? +
The browser widget uses a
zhipu-tokenizer-estimate chars-per-token estimate for English text. It is useful for rough planning, but actual billing comes from the vendor API usage fields and can differ for Chinese, code, or mixed-language prompts.