Z.ai Coding Plan: Lite vs Pro vs Max – Pricing, Credits & Differences

The Z.ai GLM Coding Plan is a subscription for developers and coding-agent users who want access to GLM models through tools such as ZCode, Claude Code, Cursor, Cline, Codex and other supported coding environments.

The most important thing to know in 2026 is that Z.ai changed the plan on July 30 from prompt-count limits to a credits-based system. That means older articles that describe the plans mainly in prompts per five hours can be misleading for new subscribers.

Quick Answer

The current GLM Coding Plan has Lite, Pro and Max tiers priced at $18, $80 and $168 per month. Each plan has both a five-hour credit limit and a weekly credit limit. Credits are consumed according to model usage, with different multipliers for GLM-5.3, GLM-5.3-Flash and supported MCP tools.

Z.ai Coding Plan Pricing

According to the official GLM Coding Plan documentation, the current individual plans use the following credit limits.

Plan	Monthly price	5-hour credits	Weekly credits	Positioning
Lite	$18	2,000	10,000	Lightweight work on small repositories
Pro	$80	12,000	60,000	Day-to-day work on mid-sized repositories
Max	$168	28,000	140,000	Advanced use on mid-to-large repositories

Z.ai’s sales pages have also shown discounted prices of $12.60 for Lite, $56 for Pro and $117.60 for Max. However, the research reviewed for this article did not find a clear official explanation of whether those figures are annual-billing discounts, introductory pricing or another promotion. Treat them as prices currently shown rather than permanent standard rates.

What Changed on July 30, 2026?

On July 30, 2026, Z.ai changed the GLM Coding Plan from a prompt-count system to a credits-based system. The company said the change was designed so deductions would better reflect actual model usage.

Older plan descriptions used rough prompt equivalents. Z.ai’s official notice gives historical equivalents of about 80 prompts per five hours for Lite, 400 for Pro and 1,600 for Max, with one prompt estimated to trigger roughly 15 to 20 model invocations.

Those older prompt figures are useful only for context. New subscribers should evaluate the current plans by credits, model multipliers and weekly limits.

How GLM Coding Plan Credits Work

Credits are calculated from token usage rather than simply counting each prompt as one unit. Z.ai’s current formula combines input tokens, cached input, output tokens and model-specific multipliers.

Usage type	GLM-5.3 multiplier	GLM-5.3-Flash multiplier
Input	6.9	2.3
Cached input	1.7	0.56
Output	24	8

Supported MCP tools such as Web Search, Web Reader and Zread use their own multiplier. Because Flash uses lower multipliers than GLM-5.3, the same plan can support substantially more Flash usage than flagship-model usage.

Z.ai Usage Limits: 5-Hour vs Weekly Quotas

Each plan has two important limits: a rolling five-hour credit allowance and a weekly allowance. Z.ai states that five-hour credits reset five hours after consumption, while weekly credits reset every seven days.

This means a user can have weekly credits remaining but still temporarily hit the five-hour limit after intensive use. The reverse is also possible: a fresh five-hour window does not help if the weekly allowance has already been exhausted.

What Are Z.ai Peak Hours and Off-Peak Hours?

Z.ai currently charges 50% of standard credits during off-peak periods. The documented Z.ai peak hours are Monday through Friday from 14:00 to 18:00 Singapore time (UTC+8). All other regular hours, including weekends, are off-peak and use half the standard model credits.

The company also documents a temporary all-day off-peak period from September 25 through October 7, 2026. Because promotions can expire or change, users should verify the latest rules in the official GLM Coding Plan FAQ before making a purchase decision based on a promotion.

Which Models Are Included?

The current Coding Plan centers on GLM-5.3 and GLM-5.3-Flash. Z.ai also documents automatic routing for older model names: requests for GLM-5.2 and GLM-5.1 route to GLM-5.3, while GLM-4.7 routes to GLM-5.3-Flash.

GLM-5.3-Flash receives a larger effective quota because its multipliers are lower. Our GLM-5.3-Flash guide explains the model’s multimodal features, API pricing and plan availability in more detail.

How to Check Z.ai Coding Plan Usage

ZCode now includes a dedicated Usage Stats view for Coding Plan subscribers. The Coding Plan tab shows your current five-hour and weekly quota status, GLM-5.3 and GLM-5.3-Flash token consumption, MCP tool usage and remaining quota.

ZCode can also display quota reset cards when your account receives one. Z.ai says these cards can restore an eligible five-hour or weekly quota to 100%, but they can expire and are only issued under qualifying conditions. The exact reset-card window and usage threshold are dynamic, so use the information shown inside ZCode rather than assuming a fixed schedule.

ZCode’s official Usage Stats documentation explains the current quota dashboard, model usage, tool usage and reset-card behavior.

Which Coding Tools Work With the Plan?

Z.ai currently lists more than 20 supported programming and general-purpose tools. Coding tools include ZCode, Claude Code, Claude for IDE, Codex, OpenCode, Cursor, Cline, TRAE, Qoder, Droid, Kilo Code, Roo Code, Crush, Goose and Eigent.

ZCode is Z.ai’s own Agentic Development Environment and currently carries a listed 1.5× usage treatment. For a broader explanation of the app itself, see our What Is ZCode? guide.

What MCP Tools Are Included?

Current plan documentation lists several MCP servers available across plans, including Vision Understanding, Web Search, Web Reader and Zread. These tools let compatible coding agents pull in additional information or process other types of content without requiring the user to build every integration manually.

Lite vs Pro vs Max: Which Plan Fits Which User?

Lite

Lite is the entry plan. Its 2,000 five-hour credits and 10,000 weekly credits make it better suited to lighter coding workloads, smaller repositories, experimentation or users who do not keep an agent running heavily throughout the day.

Pro

Pro raises the limits substantially to 12,000 credits per five-hour window and 60,000 per week. Z.ai positions it for regular development on mid-sized repositories and also advertises priority access, curated MCP tools and faster generation.

Max

Max is designed for heavier use, with 28,000 five-hour credits and 140,000 weekly credits. Z.ai positions it for advanced users working on mid-to-large repositories and gives it first access to some new flagship models and features plus dedicated resources during peak periods.

The right tier depends less on repository size alone and more on how often you use agents, which model you choose, how much cached context is reused and whether you work during off-peak periods.

Is There a Free Trial?

The source material reviewed does not document a standard free trial for the Coding Plan itself. However, ZCode offers a five-day trial for new users, and Z.ai also provides some free API models outside the subscription.

That distinction matters: a ZCode trial is not the same as a free Coding Plan subscription.

What About ZCode’s Five-Day Trial?

ZCode currently documents a five-day new-user trial that includes GLM-5.3 and GLM-5-Turbo allowances. This is useful for testing the ZCode environment before deciding whether to pay for ongoing model usage.

Can You Use the Plan With Claude Code?

Yes. Claude Code appears in Z.ai’s current supported-tools list. Z.ai also documents Anthropic-compatible and OpenAI-compatible API routes, which makes the plan relevant to developers who want to use GLM models through existing coding-agent workflows.

For a broader view of the Z.ai ecosystem, including chat, APIs and models, see our What Is Z.ai? overview.

Refunds, Cancellation and Plan Changes

Z.ai’s current FAQ says subscriptions are non-refundable once purchased and recommends canceling at least 24 hours before renewal. Same-tier plan changes can extend validity, while cross-tier upgrades use an immediate pro-rated balance conversion.

Current Promotions to Watch

As of October 2, 2026, Z.ai still documents a GLM-5.3-Flash campaign running through October 7. During specified hours, paid users can receive zero-quota Flash usage through ZCode, with increased Flash quota in other supported agents.

This is temporary promotional information, not a permanent plan feature. If you are reading this after October 7, check current Z.ai documentation before assuming the campaign is still active.

Frequently Asked Questions

Bottom Line

The Z.ai Coding Plan is best understood as a credits-based subscription for AI coding-agent usage, not a simple unlimited plan. Lite, Pro and Max differ substantially in both five-hour and weekly allowances, and effective usage also changes based on the model, cache behavior and off-peak timing.

For users evaluating the wider Z.ai ecosystem rather than only the subscription, our Z.ai overview explains how the Coding Plan fits alongside ZCode, the API platform and current GLM models.