All GLM Coding Plan Prices and Discounts
GLM Coding Plan Lite – Yearly (30% Off)
Billed $151.2 per year, which works out to $12.6 per month instead of $18. Includes 10,000 credits per week for small-repo work in Claude Code, ZCode and 20+ other tools.
GLM Coding Plan Pro – Annual Billing, Save 30%
Pay $672 for a full year of Pro, or $56 a month instead of $80. You get 6x Lite usage with 60,000 weekly credits, priority model access and faster generation.
GLM Coding Plan Max – Yearly Plan at 30% Off
One yearly payment of $1,411.2 (about $117.6 per month, down from $168). Built for mid-to-large repos with 14x Lite usage and dedicated resources during peak times.
GLM Coding Plan Lite – Quarterly (20% Off)
Billed $43.2 every three months, or $14.4 per month. A shorter commitment for testing GLM-5.3 on a single small project.
GLM Coding Plan Pro – Every 3 Months, 20% Off
Pro for $192 per quarter, which comes to $64 a month. Same 60,000 weekly credits and curated MCP tools as the yearly plan.
GLM Coding Plan Max – Quarterly Billing Deal
Max billed $403.2 per quarter, or $134.4 per month equivalent. Offers 140,000 weekly credits and first access to new flagship models.
GLM Coding Team Standard Seat – Annual (10% Off)
$950.4 per seat for a year, or $79.2 per seat per month instead of $88. Each seat gets 66,000 weekly credits, with a 2-seat minimum and centralized billing.
GLM Coding Team Premium Seat – Save 10% Yearly
$2,030.4 per seat per year, which is $169.2 a month instead of $188. Each seat includes 155,000 weekly credits plus priority resource allocation during peak hours.
How to Get the GLM Coding Plan Discount
- Click any Get Deal button above to open the Z.ai subscription page, then register or log in to your Z.ai account.
- Switch billing to Yearly (30% off) or Quarterly (20% off).
- Pick Lite, Pro or Max (or Standard/Premium seats for teams) and complete payment.
- In the Z.ai console, create an API key for the plan.
Our Take on the GLM Coding Plan
The value of this AI coding is hard to argue with. Even Lite’s 10,000 weekly credits cover an estimated 48–97 million GLM-5.3 tokens a week, and long agent sessions that reuse context stretch that further. Peak hours (14:00–18:00 Singapore time) fall overnight in the US, so American developers mostly pay the cheaper off-peak rate. Quality is solid but not top tier: GLM-5.3 still trails GPT-5.6 Sol and Claude Opus 5 on Terminal-Bench 3.0. Treat it as a strong, cheap workhorse, not a frontier replacement.
Key features
- GLM-5.3 flagship (1M context, 128K max output, always-on reasoning at low/high/max effort) plus multimodal GLM-5.3-Flash on every tier.
- Works in 20+ tools including Claude Code, Cursor, Cline, Codex, OpenCode, Roo Code, Kilo Code and Z.ai‘s own ZCode.
- Vision Understanding, Web Search, Web Reader and Zread MCP servers included on all plans.
- Credits reset on a 5-hour window and a 7-day window; hitting the cap pauses you and never charges your account balance.
- Same-tier upgrades stack validity (Lite monthly to Lite annual gives 13 months); cross-tier upgrades apply immediately with pro-rated credit.
What we think
Lite vs Pro vs Max: Which Plan Fits Your Repo?
| Lite | Pro | Max | |
|---|---|---|---|
| Yearly price | $12.60/mo | $56/mo | $117.60/mo |
| Weekly credits | 10,000 | 60,000 | 140,000 |
| 5-hour credits | 2,000 | 12,000 | 28,000 |
| Usage vs Lite | 1× | 6× | 14× |
| Recommended concurrent projects | 1 | 1–2 | 2+ |
| Est. GLM-5.3 tokens/week (95% cache) | 48–97M | 290–580M | 676–1,352M |
| Model access | Rolling access to latest | Priority access, faster speeds | First access, dedicated peak resources |
| Best for | Small repos, part-time use | Daily work on a mid-sized repo | Heavy agents on mid-to-large repos |
For teams, the Standard Seat (66,000 weekly credits) and Premium Seat (155,000 weekly credits) cannot be mixed, and Standard seats cannot be upgraded to Premium mid-subscription.
Frequently Asked Questions
Does the GLM Coding Plan work with Claude Code?
Yes. Set ANTHROPIC_BASE_URL to https://api.z.ai/api/anthropic, add your Z.ai key as ANTHROPIC_AUTH_TOKEN in ~/.claude/settings.json, and map the model slots to glm-5.3[1m].
What happens when I run out of credits?
Your tool stops getting responses until the 5-hour window resets, and Z.ai does not charge your account balance. If your weekly credits are gone, you wait for the 7-day reset.
When should I cancel to avoid auto-renewal?
Cancel at least 3 days before your next billing date to be safe.
Can I upgrade from Lite to Pro later?
Yes. A cross-tier upgrade takes effect immediately, and the unused value of your current plan converts to account balance toward the price difference. Moving from monthly to annual on the same tier adds time instead.
System Requirements & Details
| Requirement | Details |
|---|---|
| Platform | Cloud service; use via supported tools on macOS, Windows, Linux or WSL |
| Account | Z.ai account with an active GLM Coding Plan and plan API key |
| Supported tools | ZCode, Claude Code, Claude for IDE, Codex, OpenCode, Pi, Cursor, Cline, TRAE, Qoder, Droid, Kilo Code, Roo Code, Crush, Goose, Eigent (plus best-effort agents such as OpenClaw) |
| Claude Code | Latest version recommended; required for the [1m] model suffix |
| OpenClaw | Node.js 22 or newer |
| Anthropic-compatible endpoint | https://api.z.ai/api/anthropic |
| OpenAI-compatible endpoint | https://api.z.ai/api/coding/paas/v4 |
| Codex endpoint | https://api.z.ai/api/v1 |
| Models | glm-5.3, glm-5.3-flash |
| Payment | Bank card or PayPal |
