Global AI Token Comparison Hub
CN
Back to all plans

智谱 GLM

MCP tools included

GLM Coding Plan personal from ¥49/mo, team from ¥598/seat, with MCP and GLM-5.2

Coding PlanGLM Coding Plan
SubscriptionToken API
Official

Core models

GLM-5.2GLM-5-TurboGLM-4.7GLM-4.5-Air
GLM-5.2

Current flagship with 1M context and open-source SOTA coding—more stable long-horizon execution; higher quota multiplier at peak; first choice for complex engineering and agentic coding.

GLM-5-Turbo

Agent-optimized “lobster-enhanced” base with strong long-task continuity—tiered like GLM-5.2; best for long chains, especially off-peak or during the limited 1× promo window.

GLM-4.7

Mature high-intelligence model for balanced chat, reasoning, and coding at 1× quota all day—the recommended daily workhorse so not every task burns 5.2-tier quota.

GLM-4.5-Air

High-value model strong on reasoning, coding, and agents at 1× quota—good for lighter tasks, fast iteration, and pacing plan consumption.

Plan details

Lite

Lite
Price
¥49/mo
~80 prompts/5h · ~400 prompts/week
Auto monthly
¥49/mo
Auto quarterly
¥44.1/mo
Quarterly 10% off
Auto annual
¥39.2/mo
Annual 20% off
Official
Usage
Personal entry tier with ~80 prompts per 5 hours and ~400 per week—fits small repos and light iteration, not sustained heavy multi-project work.
Models
Full model list: GLM-5.2, GLM-5-Turbo, GLM-4.7, GLM-4.5-Air—use GLM-4.7 daily and switch to 5.2 for hard checkpoints.
Highlights
Web search, page read, and open-repo MCP share 100 calls/month; vision MCP shares the 5-hour prompt pool with models.
Works with Claude Code, OpenClaw, Kilo Code, and 20+ supported tools—good for individuals testing whether GLM + MCP fits their workflow.
Best for
Individual developers, light-to-medium AI coding users, and people trying an MCP-driven workflow first

Pro

Most popularRecommended
Price
¥149/mo
~400 prompts/5h · ~2000 prompts/week
Auto monthly
¥149/mo
Auto quarterly
¥134.1/mo
Quarterly 10% off
Auto annual
¥119.2/mo
Annual 20% off
Official
Usage
About 5× Lite quota: ~400 prompts per 5 hours and ~2000 per week—suitable as a daily primary coding subscription for most individuals.
Models
Earlier access to new flagship features with faster generation; 1000 combined MCP calls/month for serious search, reading, and repo tooling.
Highlights
Higher concurrency than Lite (Max > Pro > Lite); official guidance suggests Pro fits 1–2 parallel project-level agent workflows.
If you already code heavily with AI tools and want a stable search-read-edit loop, Pro is usually the most balanced personal tier.
Best for
Frequent individual developers, small teams, and users integrating MCP toolchains into daily work

Max

Heavy
Price
¥469/mo
~1600 prompts/5h · ~8000 prompts/week
Auto monthly
¥469/mo
Auto quarterly
¥422.1/mo
Quarterly 10% off
Auto annual
¥375.2/mo
Annual 20% off
Official
Usage
Personal heavy tier: ~1600 prompts per 5 hours and ~8000 per week—about 20× Lite—for deep work on larger repos and parallel projects.
Models
First access to new flagship models with peak-time priority; 4000 combined MCP calls/month to use the model as a search-read-execute workbench.
Highlights
Fits long high-frequency coding with complex task chains and heavy tool use—the value is sustained pressure without constant quota anxiety.
If you often work on large repos, complex refactors, and long-context debugging, the practical gap versus Lite/Pro becomes very noticeable.
Best for
Heavy AI coding users, large project teams, and people running high-frequency MCP toolchain workflows

团队标准版

Per seat
Price
¥598/mo
2 seats min · 60M tokens/5h/seat · 300M tokens/week/seat
Auto monthly
¥598/mo/seat
Auto annual
¥538.20/mo/seat
Annual 10% off
Official
Usage
Team billing is per seat with independent 5-hour and weekly token quotas—not one shared pool; Standard offers up to ~60M tokens/5h and ~300M tokens/week per seat.
Models
Includes seat/permission management, usage dashboards, consolidated billing/invoicing, and data not used for training by default; overage at 90% of API list price with per-member caps.
Highlights
Same models: GLM-5.2, GLM-5-Turbo, GLM-4.7, GLM-4.5-Air; 1000 combined MCP calls per seat/month (search/read/open-repo shared).
Suited to teams doing daily work on medium repos; requires a team plan key (separate from personal keys); the purchasing admin does not occupy a seat by default.
Best for
Mid-size R&D teams needing seat governance and unified billing (2 seats minimum)

团队高级版

Per seat
Price
¥1198/mo
160M tokens/5h/seat · 800M tokens/week/seat
Auto monthly
¥1198/mo/seat
Auto annual
¥1078.20/mo/seat
Annual 10% off
Official
Usage
Advanced offers up to ~160M tokens/5h and ~800M tokens/week per seat—about 2.5–2.7× Standard—for deep work on larger repos.
Models
Adds first access to new flagship models and peak-time resource priority on top of Standard governance, billing, data isolation, and overage.
Highlights
4000 combined MCP calls per seat/month; official guidance supports 2+ parallel project-level workflows with higher off-peak concurrency.
Cannot mix Standard and Advanced seats or upgrade Standard to Advanced mid-term; adding seats bills pro-rata for the remaining period.
Best for
Enterprise teams with high-concurrency, multi-project work needing peak-time guarantees

Quota

Product type
Coding Plan
Official product line
GLM Coding Plan
5-hour limit
Lite ~80 prompts · Team Standard 60M tokens/seat
Weekly limit
Lite ~400 prompts · Team Standard 300M tokens/seat
Monthly limit
Monthly MCP calls reset by tier/seat

Notes

  • Personal list: Lite ¥49, Pro ¥149, Max ¥469/mo; quarterly 10% off, annual 20% off (Lite annual ≈ ¥39.2/mo). Team Standard ¥598/seat/mo, Advanced ¥1198/seat/mo, 2 seats minimum.
  • GLM-5.2/GLM-5-Turbo are always 3× at peak (14:00–18:00 UTC+8); off-peak is normally 2×. Limited promo bills off-peak at 1× through end of September; from October expect 2× off-peak unless extended. GLM-4.7 and GLM-4.5-Air stay 1× all day. Legacy GLM-5.1/GLM-5 calls auto-route to GLM-5.2.
  • New personal purchases may show sold out: official restock daily at 10:00 UTC+8 with equal Lite/Pro/Max quotas; active auto-renewal or in-term upgrades are exempt. Team plans are not subject to this cap.
  • Plans work only in supported tools with Coding base URLs (e.g. Claude Code: `https://open.bigmodel.cn/api/anthropic`); seat sharing or unsupported tools may trigger risk controls.
  • Team plans can enable overage at 90% of API list price; personal plans pause until the next 5-hour/weekly reset and do not auto-charge account balance.

Supported coding tools

Claude CodeOpenClawOpenCodeClineCursorKilo CodeRoo CodeCrushGooseFactory DroidGemini CLIGrok CLICherry StudioTRAECodeBuddy

Pricing and model data sourced from official vendor websites

FAQ

General·8
General
8 条