GLM Coding Plan
GLM Coding Plan is Zhipu's subscription for using GLM models inside coding tools. Rechecked 2026-08-17: z.ai/subscribe sells Lite $10 / month ($8 billed yearly), Pro $30 / month ($24 yearly), and Max $60 / month ($48 yearly). Extra usage packs are $10 / 100 on Lite and $30 / 300 on Pro and Max. The old page said $3 / month unlimited GLM-4.7. That SKU is not on the table we opened.
Compare the model pages GLM-5.3 and GLM-5.2 if you wanted API weights, and Claude Code if you wanted Anthropic's own agent instead of a Zhipu seat.
Key Features
- Seat for tools, not a new IDE: The subscribe page is a model plan you plug into Claude Code, Cline, Cursor, OpenClaw, and other listed clients.
- GLM-5 line: The live compare table includes GLM-5, not a frozen GLM-4.7 catalog. Recheck which SKU your client actually sends.
- Quota plus top-up: Lite / Pro / Max are monthly pools. Extra packs sit on the same page.
- Not MIT weights: The Coding Plan is a hosted subscription. Open weights live on the model pages.
Use Cases
- People who already live in Claude Code or Cline and want a cheaper GLM seat than Anthropic's own plan.
- Teams comparing Lite $10 vs Pro $30 after they see weekly quota burn.
- Users who still have a $3 GLM-4.7 bookmark and need the current table.
Limitation: "Unlimited" and the LiveCodeBench / SWE-bench percentages from the old page are out. We did not re-verify first-month discounts or token-per-week math beyond the public rows.
Pricing
z.ai/subscribe on 2026-08-17. Recurring monthly first.
| Plan | Recurring | Extra pack | Notes from that page |
|---|---|---|---|
| Lite | $10 / month ($8 yearly) | $10 / 100 | Entry seat. |
| Pro | $30 / month ($24 yearly) | $30 / 300 | Standard seat. |
| Max | $60 / month ($48 yearly) | $30 / 300 | Highest public row. |
If a leftover blog still says $3 / ¥20 unlimited, ignore it.
Getting Started
- Open z.ai/subscribe and start on Lite only long enough to see weekly quota.
- Point Claude Code, Cline, or Cursor at the Zhipu endpoint before you buy Max.
- Recheck GLM-5.3 if you wanted API or weights instead of this seat.
- Do not paste 84.9% LiveCodeBench into a deck.
Frequently Asked Questions
Is it still $3 unlimited?
No. The live public rows are Lite $10, Pro $30, and Max $60.
Is the model still GLM-4.7?
The subscribe table we opened includes GLM-5. Use the model pages for the current flagship name.
Same as Claude Code?
No. This is a Zhipu model subscription. Claude Code is Anthropic's agent. You can point some agents at this plan.
Alternatives
- Claude Code: Anthropic's own agent across terminal, IDE, and desktop.
- Cursor: AI IDE with its own $20 / $40 seats.
- GLM-5.3: Current Zhipu flagship model page, not this subscription.
Tips
- Quote $10 / $30 / $60 from the recurring column.
- Do not call it unlimited GLM-4.7.
- Recheck extra-pack prices before you assume overage is cheap.
Conclusion
GLM Coding Plan is a priced Zhipu seat for coding tools, not a $3 unlimited GLM-4.7 SKU. Start at z.ai/subscribe, then decide whether Claude Code or GLM-5.3 already covers the access you need.
Comments
No comments yet. Be the first to comment!
Related Tools
TRAE
www.trae.ai
ByteDance's AI IDE. Recurring seats: Lite $3, Pro $10, Pro+ $30, Ultra $100. Usage is billed in dollars, not extra-fast request packs.
Qoder
qoder.com
Alibaba's AI IDE. Recurring seats: Lite $10, Pro $20, Pro+ $60, Ultra $200. Usage is Auto credits.
Factory Droid
factory.ai
Factory coding agents. Rechecked: Pro $20 / Plus $100 / Max $200. Not a 58.75% Terminal-Bench census.
Related Insights
Skills + Hooks + Plugins: How Anthropic Redefined AI Coding Tool Extensibility
An in-depth analysis of Claude Code's trinity architecture of Skills, Hooks, and Plugins. Explore why this design is more advanced than GitHub Copilot and Cursor, and how it redefines AI coding tool extensibility through open standards.
Seven AI Coding CLIs, Six Months: No Matter How Strong the Model, Work Needs Supervision
Claude Code, Codex, opencode, pi, omp and DeepSeek Harness all have personalities. After six months of deep use I run a division of labor: pi for the fastest cheapest reviews, omp for complex PRs, DeepSeek Harness on V4 Flash for high-frequency low-cost review, and Claude Code, Qoder and Cursor for writing. No matter how strong the model, work needs supervision — ideally from an independent third party.
Hook OpenCode Go into Codex on Windows. Do Not Open a Second Toolkit.
A ChatGPT-signed Codex desktop app still shows mostly GPT in the picker. On Windows, enable only OpenCode Go and the Grok, GLM, Kimi, DeepSeek, and MiniMax models you already pay for appear in the same selector. Keys stay local. Native GPT stays put.