GLM 5.3 Price: API Cost, Coding Plan Pricing & Free Options
Quick answer: GLM 5.3 has no separate price tag. It ships inside the GLM Coding Plan — Lite $18/mo, Pro $80/mo, Max $168/mo (quarterly −20%, yearly −30%), now with a points-based quota where off-peak hours cost 50% of standard points. The public API is coming soon and is expected to follow GLM 5.2's rates ($1.4 per 1M input / $4.4 per 1M output). The truly free option: open weights on HuggingFace in two weeks.
TL;DR
| Access | Price | Notes |
|---|---|---|
| Coding Plan Lite | $18/mo | 10,000 credits/week |
| Coding Plan Pro | $80/mo | 6× Lite usage |
| Coding Plan Max | $168/mo | 14× Lite usage, priority |
API (glm-5.3) | Coming soon | Expected at GLM 5.2 rates: $1.4/$4.4 per 1M |
| Open weights | Free | HuggingFace in ~2 weeks |
| Off-peak discount | 50% points | Outside 14:00–18:00 UTC+8 weekdays |
GLM 5.3 Is Not Priced Separately
Z.AI rolled GLM 5.3 out to all GLM Coding Plan subscribers on launch day (August 14, 2026) — there is no "5.3 premium" or per-model surcharge. Your plan price is your 5.3 price.
The new plan structure uses a points-based quota: input, cached input, and output tokens each consume points at their own rate. The single biggest cost lever: calls outside peak hours (14:00–18:00 UTC+8, Mon–Fri) consume only 50% of standard points — including all weekend.
GLM Coding Plan Pricing (2026)
| Plan | Monthly | Quarterly (−20%) | Yearly (−30%) | Quota |
|---|---|---|---|---|
| Lite | $18 | $14.4/mo | $12.6/mo | 10,000 credits/week |
| Pro (Popular) | $80 | $64/mo | $56/mo | 6× Lite |
| Max | $168 | $134.4/mo | $117.6/mo | 14× Lite + priority |
What you get at every tier:
- Rolling access to the latest flagship models — GLM 5.3 included, automatically.
- 20+ agent tools — ZCode, Claude Code, Cline, Kilo Code, OpenCode and more.
- Lite: built for lightweight iteration on small repos.
- Pro: 6× Lite usage, priority access to latest models, curated MCP tools, faster generation — for day-to-day dev on mid-sized repos.
- Max: 14× Lite usage, first access to new models, dedicated resources during peak times — for mid-to-large repos.
Annual billing at −30% is the cheapest way in: Lite works out to ~$0.42/day.
Expected API Pricing
The glm-5.3 API endpoint is coming soon — the model is currently available through Coding Plan and ZCode only. Z.AI's official pricing page has not yet listed a separate GLM-5.3 rate, and given the same base model, pricing is expected to mirror GLM 5.2:
| Model | Input / 1M | Cached input / 1M | Output / 1M |
|---|---|---|---|
| GLM-5.2 (official) | $1.4 | $0.26 | $4.4 |
| GLM-5.1 | $1.4 | $0.26 | $4.4 |
| GLM-5 | $1.0 | $0.20 | $3.2 |
| GLM-5-Turbo | $1.2 | $0.24 | $4.0 |
For reference, that makes GLM 5.2-class pricing roughly 4× cheaper than Claude Opus 5 ($5 input / $25 output) at comparable open-weights capability — and GLM 5.3 is more token-efficient than 5.2 (34.5% completion at ~75K tokens vs 23.4% at ~96K on Z.AI's code bench), so effective cost per completed task drops further.
The Free Option: Open Weights
GLM 5.3 weights land on HuggingFace two weeks after launch (safety evaluation permitting). Self-hosting eliminates per-token cost entirely — the same playbook already used for GLM 5.2. If your workloads are high-volume, predictable, or privacy-sensitive, wait for the weights and run your own cluster.
How to Minimize GLM 5.3 Cost
- Schedule off-peak. 50% points outside 14:00–18:00 UTC+8 weekdays — evenings and weekends are half price.
- Cache aggressively. Stable system prompts and context prefixes hit the cached-input rate ($0.26/1M on the API; ~30% more effective tokens in ZCode).
- Right-size effort.
reasoning_effort: lowfor extraction/classification,maxonly for complex coding — effort drives thinking tokens. - Use ZCode before Aug 31. The 1.5× limited-time quota boost stacks with cache savings (up to 180% of standard quota).
- Annual billing. −30% on any Coding Plan tier.
FAQ
How much does GLM 5.3 cost? There's no separate 5.3 price. It's included in the GLM Coding Plan ($18/$80/$168 per month) and, when the API launches, expected to bill at GLM 5.2 rates ($1.4 input / $4.4 output per 1M tokens).
Is GLM 5.3 free? Coding Plan subscribers get it within their plan; off-peak usage is 50% of standard points. Open weights (free, self-hosted) arrive in about two weeks.
What are GLM Coding Plan prices? Lite $18/mo, Pro $80/mo, Max $168/mo; quarterly −20%, yearly −30% (Lite $12.6/mo on annual).
Is the GLM 5.3 API available yet? Not yet — Z.AI says it's coming soon. Coding Plan and ZCode are the current access paths.
Is GLM 5.3 cheaper than Claude Opus 5? At expected 5.2-level pricing ($1.4/$4.4), GLM 5.3 is roughly 4× cheaper per token than Opus 5 ($5/$25) — and it's open-weights.
Sources
- Z.AI: GLM-5.3: Frontier Coding with Emergent Cyber Capabilities (August 14, 2026)
- Z.AI Coding Plan
- Z.AI Docs: Pricing
Last updated: August 14, 2026