Quick answer: On Z.ai's official API, GLM-5 costs $1.00 input / $3.20 output per 1M tokens, GLM-5.2 and GLM-5.3 cost $1.40 / $4.40, GLM-5.3-Flash costs $0.15 / $0.50, and GLM-5.3-FlashX costs $0.37 / $1.25. If you mostly code inside tools like Claude Code or OpenCode, Z.ai's GLM Coding Plan ($18, $80 or $168 a month) is usually cheaper than paying per token. If you want GLM 5.3 chat plus an API key without building anything, a third-party plan such as glm5.app's is the third option. This guide shows when each one costs least.
We checked every price here on November 5, 2026, using Z.ai's pricing page, the GLM Coding Plan docs and subscription page, the GLM-5 model page, OpenRouter's live provider list, and glm5.app's own pricing and billing docs. glm5.app is the site you are reading. It is an independent service, not Z.ai, and the sections below say where Z.ai's own offer is the better buy.
Who this guide is for
"GLM 5 pricing" sounds like one number, but the family now has five API models, two subscription products and dozens of resellers. Most pricing pages cover one model, so you have to work out for yourself which number applies to your workload. This page compares the whole family on one set of requests. Use it if you are choosing between:
- the official per-token API,
- the GLM Coding Plan subscription,
- a third-party route (OpenRouter providers or glm5.app).
For single-model detail, see the dedicated guides on GLM 5.3 pricing, GLM 5.2 pricing and GLM 5.3 Flash pricing.
Official GLM-5 family API prices (Z.ai)
Z.ai's pricing page lists these rates in USD per 1 million tokens:
| Model | Input | Cached input | Output | Notes |
|---|---|---|---|---|
| GLM-5.3 | $1.40 | $0.26 | $4.40 | Current flagship; 1M context |
| GLM-5.3-Flash | $0.15 | $0.03 | $0.50 | Cheapest current model |
| GLM-5.3-FlashX | $0.37 | $0.075 | $1.25 | Faster serving tier of Flash |
| GLM-5.2 | $1.40 | $0.26 | $4.40 | Same rate card as GLM-5.3 |
| GLM-5 | $1.00 | $0.20 | $3.20 | Original GLM-5; 200K context |
Three things to know before you use this table:
- Cached-input storage is "Limited-time Free" for every model on the page. Cache hits are billed at the cached-input rate, and storage may be charged later.
- GLM-5 is the cheapest large model but the oldest. Z.ai's model page lists a 200K context window and 128K max output for GLM-5, and describes it as 744B total parameters with 40B active. GLM-5.3 lists a 1M context window.
- Built-in Web Search costs $0.01 per use, separate from tokens. Agents that search often need to budget for it.
Z.ai's open weights for GLM-5 were published on Hugging Face under zai-org/GLM-5 on February 11, 2026, under the MIT license. That explains the "release date pricing performance" searches: the price has not changed since launch, while newer models have arrived above and below it.
The GLM Coding Plan (Z.ai's subscription)
When people search "GLM 5 subscription" or "GLM5 coding plan", they usually mean this product. It is billed in credits, not tokens:
| Tier | Monthly | Yearly (per month) | 5-hour credits | Weekly credits |
|---|---|---|---|---|
| Lite | $18 | $12.60 | 2,000 | 10,000 |
| Pro | $80 | $56 | 12,000 | 60,000 |
| Max | $168 | $117.60 | 28,000 | 140,000 |
According to Z.ai's Coding Plan docs:
- Supported models are GLM-5.3 and GLM-5.3-Flash. Requests for GLM-5.2 or GLM-5.1 are automatically routed to GLM-5.3. GLM-5 and FlashX are not on the supported list.
- Credits per request = (input × 6.9 + cached input × 1.7 + output × 24) ÷ 10,000 for GLM-5.3. The multipliers for Flash are 2.3 / 0.56 / 8.
- Off-peak usage costs 50% of the normal credits. Peak hours are Monday–Friday, 14:00–18:00 Singapore time.
- It works in agent tools such as ZCode, Claude Code, Codex, OpenCode and Cline. It is not a general API balance, and Coding Plan keys follow their own endpoint rules.
For the full tier breakdown, see our GLM 5.3 Coding Plan guide.
glm5.app plans, packs and API rates
glm5.app sells one credit balance. You can spend it on web chat, image and video generators, and an OpenAI-compatible API at https://glm5.app/api/v1.
Subscriptions (unlimited chat is subject to fair use):
| Plan | Monthly | Yearly (per month, billed yearly) | Credits | Unlimited chat on |
|---|---|---|---|---|
| Starter | $19.9 | $9.9 | 1,500/mo or 18,000/yr | GLM 5.3 Flash |
| Plus | $29.9 | $14.9 | 2,500/mo or 30,000/yr | GLM 5.3 and GLM 5.3 Flash |
| Enterprise | $79.9 | $39.9 | 6,000/mo or 75,000/yr | GLM 5.3 and budget models |
Pay-as-you-go packs never expire and are a one-time payment. They are also the only option that accepts Alipay and WeChat Pay: Plus Pack $12 (550 credits), Pro Pack $20 (1,200), Enterprise Pack $50 (3,250), Power Pack $100 (7,000), Scale Pack $200 (15,000).
API rates per 1M tokens (input / output):
| Model ID | glm5.app API | Official Z.ai | Verdict |
|---|---|---|---|
glm-5.3 | $1.40 / $4.40 | $1.40 / $4.40 | Same list rate |
glm-5.2 | $1.40 / $4.40 | $1.40 / $4.40 | Same list rate |
glm-5.3-flash | $0.075 / $0.25 | $0.15 / $0.50 | Half the official list rate |
glm-5.3-flashx | $1.85 / $6.25 | $0.37 / $1.25 | Use Z.ai for FlashX |
glm-5 | $1.50 / $5.00 | $1.00 / $3.20 | Use Z.ai for GLM-5 |
How API usage turns into credits, per our billing docs:
- 100 credits equal $1.80 of the list usage above.
- Each completed request is rounded up to whole credits, with a minimum of 1.
- glm5.app has no separate cached-input discount.
- API access unlocks after your first purchase. Free welcome credits only work in the web app.
What you actually pay per credit depends on how you buy it. Small packs work out to about $0.022 per credit, the Scale Pack and monthly plans to about $0.012–$0.013, and yearly plans to about $0.006–$0.0066. That spread is the main reason the worked examples below give ranges.
Worked example 1: one 20K-in / 2K-out coding request
This is a typical single agent step: about 20,000 tokens of code and context in, about 2,000 tokens out, with no cache hits.
| Model | Official API | Coding Plan credits (peak / off-peak) | glm5.app credits → effective $ |
|---|---|---|---|
| GLM-5.3 | $0.0368 | 18.6 / 9.3 | 3 credits → $0.018–$0.065 |
| GLM-5.2 | $0.0368 | routed to GLM-5.3 | 3 credits → $0.018–$0.065 |
| GLM-5 | $0.0264 | not supported | 3 credits → $0.018–$0.065 |
| GLM-5.3-FlashX | $0.0099 | not supported | 3 credits → $0.018–$0.065 |
| GLM-5.3-Flash | $0.0040 | 6.2 / 3.1 | 1 credit → $0.006–$0.022 |
What this shows:
- Coding Plan Lite gives you 10,000 weekly credits. That is roughly 537 of these GLM-5.3 requests a week at peak and about twice as many off-peak, for $18 a month. Paying per token, 537 requests would cost about $19.80 a week. For daily coding inside agent tools, the Coding Plan is the cheapest route.
- On glm5.app, this GLM-5.3 request costs less than official on a yearly plan (about $0.018–$0.02 versus $0.0368), roughly the same on monthly plans and the Scale Pack (about $0.036–$0.04), and more on the smallest pack (about $0.065). Small requests lose some value to whole-credit rounding.
- For small Flash and FlashX calls, Z.ai's per-token API is cheaper.
Worked example 2: a large 200K-in / 10K-out request
When a request is large, rounding up to a whole credit hardly matters:
| Model | Official API | glm5.app |
|---|---|---|
| GLM-5.3 | $0.324 | 18 credits → $0.108 (yearly) to $0.393 (Plus Pack) |
| GLM-5.3-Flash | $0.035 | 1 credit ($0.0175 of usage) → $0.006–$0.022 |
glm5.app's half-price Flash rate matters most for big-context Flash jobs such as long-document summarization or repository Q&A.
Worked example 3: a month of chat
Assume 30 chat turns a day for 30 days, with about 8K input tokens (including history) and 1K output tokens per turn. That is 7.2M input and 0.9M output tokens:
| Route | Monthly cost |
|---|---|
| Official API, GLM-5.3 | about $14.04 + the cost of building or running your own client |
| Official API, GLM-5.3-Flash | about $1.53 + your own client |
| glm5.app Plus | $29.9 monthly or $14.9/month yearly. Unlimited GLM 5.3 chat (fair use) plus 2,500 credits for API calls and generators |
| glm5.app Starter | $19.9 monthly or $9.9/month yearly. Unlimited GLM 5.3 Flash chat (fair use) |
If you can write your own client and only chat lightly on Flash, raw tokens are cheaper. A subscription makes sense when you want a ready chat app, heavy GLM 5.3 use without counting tokens, and an API key on the same balance.
Want GLM 5.3 chat and an API key on one balance? Compare glm5.app plans and packs. Packs never expire, Alipay and WeChat Pay work on packs, and one key covers GLM 5.3, 5.2 and Flash with no Coding Plan endpoint rules.
Who is the cheapest GLM-5 API provider?
For the original GLM-5, OpenRouter's endpoint list on November 5, 2026 showed StreamLake and GMICloud as the cheapest at $0.60 / $1.92 per 1M tokens. Both serve FP8 with roughly a 198K–203K context window. Z.ai's own endpoint was $1.00 / $3.20. Resellers change prices often, and some cut cost by quantizing the model or shortening the context, so check precision and context length on the OpenRouter GLM 5.3 route guide before you commit.
For GLM-5.3, the cheapest OpenRouter endpoints had low input prices but often FP4 quantization. Z.ai and most large hosts charged $1.40 / $4.40. A low headline price is worth little if the endpoint fails your evals.
What should you pick?
| If you are… | Pick | Why |
|---|---|---|
| Coding daily in Claude Code, OpenCode, Cline or ZCode | Z.ai GLM Coding Plan (Lite to start) | Credits plus 50% off-peak are far cheaper than per-token use for agent loops |
| Building a production app with repeated prompts | Official Z.ai API | Cached input at $0.26 (GLM-5.3) and $0.03 (Flash) is a real discount |
| Using GLM-5.3-FlashX or the original GLM-5 via API | Official Z.ai API | glm5.app's rates for these two are higher |
| Chasing the lowest GLM-5 token price | Cheapest OpenRouter endpoint, after testing | $0.60 / $1.92 at check time; FP8 and a shorter context |
| A chat user who wants GLM 5.3 without building anything | glm5.app Plus, yearly if you'll stay | Unlimited GLM 5.3 chat (fair use) plus credits |
| Running large-context Flash jobs through an API | glm5.app API | $0.075 / $0.25 is half Z.ai's list rate |
| Paying with Alipay or WeChat Pay, or avoiding subscriptions | glm5.app credit packs | One-time, never expire |
FAQ
How much does GLM 5 cost? On Z.ai's API, GLM-5 is $1.00 per 1M input tokens and $3.20 per 1M output tokens. The newer GLM-5.3 is $1.40 / $4.40, and GLM-5.3-Flash is $0.15 / $0.50.
Is there a GLM 5 subscription? Yes, two kinds. Z.ai's GLM Coding Plan costs $18, $80 or $168 a month and is built for coding tools. Third-party apps like glm5.app sell chat-plus-API plans from $19.9 a month, or $9.9 a month billed yearly.
Does the GLM Coding Plan include GLM-5? Z.ai's docs list GLM-5.3 and GLM-5.3-Flash as the supported models. GLM-5.2 and GLM-5.1 requests are routed to GLM-5.3. The original GLM-5 and FlashX are not listed.
Who is the cheapest GLM-5 API provider? At our November 5, 2026 check, StreamLake and GMICloud on OpenRouter listed GLM-5 at $0.60 / $1.92 per 1M tokens (FP8). Z.ai's official price is $1.00 / $3.20.
Is GLM-5.3 more expensive than GLM-5? Yes, per token: $1.40 / $4.40 versus $1.00 / $3.20. GLM-5.3 offers a 1M context window to GLM-5's 200K and is the model the Coding Plan supports.
Can I use glm5.app credits for the API right away? Only after your first purchase. Welcome credits work in the web app only. Paid credits, from packs or plans, cover both chat and API.
Bottom line
GLM 5 pricing has three tiers. Per-token API is best for products and cache-heavy apps. The Coding Plan is best for daily agentic coding. A chat-plus-API plan is best when you want GLM 5.3 ready to use with one balance. Start from your real workload, run it through the worked examples above, and check today's rates before you buy.
If the third option fits, see glm5.app pricing for plans and never-expiring packs, or try GLM 5.3 in the browser first.
Sources
Checked on November 5, 2026:
- Z.ai API pricing: official per-token rates for GLM-5, GLM-5.2, GLM-5.3, GLM-5.3-Flash and GLM-5.3-FlashX, plus Web Search pricing.
- Z.ai GLM Coding Plan overview: supported models, credit formula, multipliers, 5-hour and weekly limits, off-peak rule.
- Z.ai Coding Plan subscription page: Lite, Pro and Max monthly and yearly prices.
- Z.ai GLM-5 model guide: GLM-5 context length, max output and parameter scale.
- OpenRouter GLM-5 model page: third-party provider prices, quantization and context per endpoint.
