Z.ai GLM Coding Plan: price, limits and resets
Z.ai sells the GLM Coding Plan in three tiers, Lite, Pro and Max, and caps each one twice: a five-hour limit and a weekly limit, both counted in credits. On , Lite cost $18 a month, Pro $80 and Max $168, billed monthly; quarterly billing took 20% off and yearly billing 30%.
A credit is not a request. Z.ai turns each call into credits from the tokens read, cached and written, at a rate set per model, and halves that rate outside peak hours. Before a long task, what matters is how much of the current five hours and of the week is left, and when each comes back.
Checked on , against Z.ai's subscription page and its Coding Plan documentation, listed under Sources. Prices in US dollars, as Z.ai shows them.
GLM Coding Plan tiers and prices
| Tier | Billed monthly | Billed quarterly | Billed yearly |
|---|---|---|---|
| Lite | $18 a month | $14.40 a month | $12.60 a month |
| Pro | $80 a month | $64 a month | $56 a month |
| Max | $168 a month | $134.40 a month | $117.60 a month |
Source: z.ai/subscribe, checked on . Z.ai shows quarterly and yearly prices per month. Z.ai also has a team plan, with its own limits.
Pro gives six times Lite's usage and Max fourteen times. Every tier runs GLM-5.3 and GLM-5.3-Flash; Z.ai sends requests for GLM-5.2 and GLM-5.1 to GLM-5.3, and those for GLM-4.7 to GLM-5.3-Flash. Z.ai does not refund a subscription once it is bought, even unused, so the billing period is worth choosing with care.
The five-hour and weekly limits
| Tier | Credits per five hours | Credits per week |
|---|---|---|
| Lite | 2,000 | 10,000 |
| Pro | 12,000 | 60,000 |
| Max | 28,000 | 140,000 |
A model call costs its input tokens times the input rate, plus its cached input tokens times the cached rate, plus its output tokens times the output rate, divided by 10,000. For GLM-5.3 the three rates are 6.9, 1.7 and 24; for GLM-5.3-Flash, 2.3, 0.56 and 8. A call to the Web Search, Web Reader or Zread tools costs 1.2 credits.
Outside peak hours, model usage costs half the credits. Peak hours run Monday to Friday, 14:00 to 18:00 Singapore time (UTC+8), that is 06:00 to 10:00 UTC. The same task can therefore use twice as much of the limit during those four hours as outside them. Z.ai publishes no monthly limit for the individual tiers.
When the GLM limits reset
- The five-hour limit refreshes on its own: Z.ai says the quota resets five hours after it is used, so the five hours run from your own use, not from a fixed hour.
- The weekly limit starts counting when you place your order and resets every seven days. Your week starts on the day you subscribed, not on Monday.
- Reset cards, handed out during some campaigns, restore the five-hour or the weekly quota at once. The window then restarts from the moment the card is used.
Once a limit is reached, calls stop until it resets. Z.ai says it does not take the shortfall from your account balance.
Where to see what is left
In Z.ai's console, the subscription page shows how far you are into the weekly quota, and the Charge Type page lists the tokens used per pricing type and the number of tool calls.
Tracking the GLM Coding Plan in Quotum
Z.ai leaves no sign-in on your Mac to read, so it is tracked with a key you paste: in Quotum, you add a Z.ai account, paste an API key and choose the account's region. For Z.ai, Quotum reads no other tool's credentials: it uses the key you paste yourself, keeps it in its own Keychain item and sends it only to Z.ai's quota server, api.z.ai, or open.bigmodel.cn for an account in China; this key can also spend with Z.ai, which Quotum never does.
Quotum then shows, one line per limit, the share used of the five-hour and weekly limits, the time each resets and how long ago it was measured, next to your Claude Code and Codex windows. The most used window sets the percentage in the menu bar. A limit Quotum cannot classify, such as a monthly one, keeps its raw name on its own line and never leads; the limit on tool calls is not shown. A key Z.ai refuses reads “Key refused”, and a key with no coding plan behind it “Measurement unavailable for this account”.
We wrote and tested this reading on , against the reply shape three third-party clients agree on; we have never observed a real reply from Z.ai. Details: What Quotum reads on your Mac.
The trial is free, lasts fourteen days and covers the first two lines of the panel. Launch price: €4.99, then €9.99. Reading a provider's quota from a third-party app is not something the provider provides for: the pricing page explains the risk before any purchase.
Other coding plans
Kimi Code and MiniMax sell coding plans capped the same way, by a short window and a longer one: see the Kimi Code plan and the MiniMax coding plan. For Claude's session and weekly limits, see Claude usage limits.
Sources
Consulted on .
- Z.ai, GLM Coding Plan:
z.ai/subscribe - Z.ai, Coding Plan overview:
docs.z.ai/devpack/overview - Z.ai, Coding Plan FAQ:
docs.z.ai/devpack/faq - Z.ai, Usage Policy:
docs.z.ai/devpack/usage-policy
FAQ
How much does the GLM Coding Plan cost?
On November 18, 2026, Lite cost $18 a month, Pro $80 and Max $168, billed monthly. Quarterly billing took 20% off and yearly billing 30%, down to $12.60, $56 and $117.60 a month.
What are the GLM Coding Plan limits?
Each tier has a five-hour limit and a weekly limit, in credits: Lite 2,000 and 10,000, Pro 12,000 and 60,000, Max 28,000 and 140,000. Outside peak hours, model usage costs half the credits.
When does the GLM Coding Plan reset?
The five-hour quota resets five hours after it is used. The weekly quota counts from the time of your order and resets every seven days.
Can I get a refund on the GLM Coding Plan?
No. Z.ai says a subscription cannot be refunded once bought, even if you have not used it up.
How can I see my GLM Coding Plan usage?
In Z.ai's console, the subscription page shows the weekly quota's progress and the Charge Type page the tokens and tool calls. Quotum shows the five-hour and weekly limits in your Mac menu bar, with a key you paste.