Limits guide
GLM Coding Plan: prices, credits, and usage limits
Z.ai's GLM Coding Plan explained: Lite, Pro, and Max prices, how 5-hour and weekly credits are counted, and how many coding sessions they hold.
· Facts checked October 3, 2026
- Claude Code
- OpenCode
- Codex
- Cursor
The short answer
The GLM Coding Plan is Z.ai's subscription for using GLM-5.3 and GLM-5.3-Flash in coding tools such as Claude Code, OpenCode, and Cline, at $18, $80, or $168 a month for Lite, Pro, and Max. Since July 30, 2026 it meters usage in credits computed from tokens, with a five-hour allowance and a weekly allowance, rather than counting prompts.
At a glance
| Prices | Lite $18, Pro $80, Max $168 a month. Quarterly billing takes 20% off, yearly 30%. |
|---|---|
| 5-hour credits | Lite 2,000, Pro 12,000, Max 28,000. Each use is restored five hours after it was spent. |
| Weekly credits | Lite 10,000, Pro 60,000, Max 140,000, resetting every seven days from the subscription start. |
| Models | GLM-5.3 and GLM-5.3-Flash. Off-peak use costs half the credits. |
| Where it works | Supported coding tools only; the plan key cannot be used for general API calls. |
| Where to check | The subscription page in Z.ai's API key console shows weekly progress. |
How credits are counted
Z.ai converts every request into credits with a per-model formula: tokens times a multiplier for each category, divided by 10,000. For GLM-5.3 the multipliers are 6.9 for uncached input, 1.7 for cached input, and 24 for output. GLM-5.3-Flash uses 2.3, 0.56, and 8, so the same work costs roughly a third of the credits.
Requests outside peak hours cost half. Z.ai defines peak as 14:00 to 18:00 Singapore time on weekdays, and weekends are off-peak all day. Older plans that counted prompts per five hours are now called Legacy plans.
What the allowances hold in practice
Applying Z.ai's formula to this site's standard agentic session, 500,000 uncached input tokens, 2 million cached tokens, and 50,000 output tokens, gives 805 credits on GLM-5.3 at peak rates. On Lite, the five-hour allowance of 2,000 credits holds about two and a half such sessions, and the weekly 10,000 about twelve. Pro holds about fifteen per five hours and seventy-four per week, and Max about thirty-four and one hundred seventy-three.
The same session priced at GLM-5.3's API rates costs $1.44, so Lite's weekly credits cover about $18 of API-equivalent use each week for a plan that costs $18 a month. Off-peak work doubles those figures, and GLM-5.3-Flash, at 267 credits for the session, stretches them about three times further. Real sessions vary, so treat these as orders of magnitude; the GLM-5.3 cost guide shows the API side in detail.
Supported tools and terms
Z.ai lists Claude Code, Codex, OpenCode, Cursor, Cline, Kilo Code, Roo Code, Crush, Goose, and its own ZCode among the supported coding agents, with endpoints for the Anthropic Messages protocol, OpenAI Chat Completions, and the OpenAI Responses API. The plan is limited to these tools: Z.ai's FAQ says API calls outside the plan are not available, and account sharing can lead to rate limiting or a ban.
To use the plan in Claude Code, follow the Claude Code setup guide. For OpenCode, or to pay per token through OpenRouter instead, see GLM in OpenCode and OpenRouter.
Watching your usage
Z.ai's console shows weekly progress on the subscription page and token use by charge type on the billing page; Z.ai documents no public API for the remaining quota. Claude Code's own usage bars and status line limits describe Claude subscriptions, so they stay empty when Claude Code runs on GLM. everyaitoken still counts the tokens from Claude Code and OpenCode history, which shows how much work went to GLM, but it does not price Z.ai's plan and labels those tokens as unpriced.
Your own numbers
See every limit and its reset time.
everyaitoken reads Claude Code, Codex, Cursor, OpenRouter, Gemini CLI, and OpenCode on your Mac. It shows limits with reset times, API-equivalent cost, and what caching saved or cost, in the menu bar and 17 widgets. $9 once.
FAQ
Questions
How much is the GLM Coding Plan?
Lite costs $18 a month, Pro $80, and Max $168. Paying quarterly takes 20% off and paying yearly takes 30% off. Z.ai does not offer refunds.
How many prompts does the GLM Coding Plan allow?
Current plans count credits rather than prompts. Lite gets 2,000 credits per five hours and 10,000 per week; Pro gets 12,000 and 60,000; Max gets 28,000 and 140,000. How many prompts that covers depends on the tokens each one uses.
Can I use the GLM Coding Plan with Claude Code?
Yes. Z.ai supports Claude Code through its Anthropic-compatible endpoint. The setup guide lists the settings.
Can I use the GLM Coding Plan key for my own app?
No. Z.ai limits the plan to its supported coding tools and says API calls outside the plan are not available. For an app, use a pay-as-you-go API key.
How does the GLM Coding Plan compare with Claude Max?
They meter differently: Z.ai publishes credit formulas, while Anthropic publishes plan multiples. The Claude Max guide explains the Claude side of the comparison.
Sources
Checked October 3, 2026. Plans and limits change; the linked pages are the authority.