Z.ai’s GLM models on a US dollar cost sheet
Compare GLM-5.3, Flash and FlashX using fresh input, cached input and output rates for an English application.

👁 111 reads
A library search box receives long queries and returns short shelf locations. A writing assistant can have the reverse shape. Z.ai lists separate rates for fresh input, cached input and generated text, so I would price those two products from their own usage logs.
| Model | Fresh input | Output |
|---|---|---|
| GLM-5.3 | $1.40 | $4.40 |
| GLM-5.3-Flash | $0.15 | $0.50 |
| GLM-5.3-FlashX | $0.37 | $1.25 |
All figures are per million tokens. The FlashX-Legal listing has its own model label even where its tariff matches FlashX. If the deployment contract calls for a particular edition, retain the model ID with the quote.
For the library I would grade exact shelf locations, including answers that should say the book is absent. A strictly cheaper answer that points the patron to the wrong floor can consume a librarian’s afternoon.
Practical details
Z.ai lists $0.26 per million cached input tokens for GLM-5.3. It labels cache storage as a limited-time free offering.
Google · Your sourcesQuantslant higher in your Google - one clickAdd to preferred sources →Sources
- https://docs.z.ai/guides/overview/pricing (accessed November 17, 2026)
More on this from QUANTSLANT
benchmarks·Xiaomi models, access and benchmark guideXiaomi model and access guide with a practical US-English evaluation brief and links to publisher materials.