GLM-5.3 API pricing
Zhipu's current flagship. Available at listed at platform rate (no published official USD price).
| Route | Input / 1M | Output / 1M |
|---|---|---|
| Platform rate | $0.7 | $2.2 |
USD, standard tier, verified 2026-09-16.
Specifications
| Vendor | Zhipu |
| Context window | — |
| Max output | — |
| Category | China |
What GLM-5.3 costs at your volume
| Usage level | Input / output per month | Monthly cost |
|---|---|---|
| Light — prototyping, a few thousand calls | 5M / 1M | $5.70 |
| Working — one developer, daily use | 50M / 10M | $57.00 |
| Heavy — team or agent loops in production | 500M / 100M | $570.00 |
Calculated at $0.7 input / $2.2 output per 1M tokens, assuming a 5:1 input-to-output ratio. Your ratio decides the real number — run your own figures through the cost calculator.
When GLM-5.3 is the right call
At $0.7 per million input tokens this sits in the mid-range band. It is a good fit for the everyday mix — agent loops, chat assistants, code editing, and anything interactive where latency and quality both matter.
It is usually the wrong call for either extreme: trivial bulk jobs that a cheaper model handles identically, or the hardest reasoning tasks where a flagship earns its premium.
Similarly priced alternatives
| Similarly priced | Vendor | In / out per 1M | vs this |
|---|---|---|---|
| Claude Haiku 4.5 | Anthropic | $0.5 / $2.5 | +$0.10 |
| DeepSeek V4 Pro | DeepSeek | $0.66 / $1.98 | −$0.26 |
| Gemini 3.7 Flash | $0.375 / $1.875 | −$0.65 |
Blended comparison (input + output per 1M). Price is one axis — the models above are not interchangeable, they are the ones worth testing side by side at this budget.
Setting it up
Pointing an existing integration at this model is a base URL change, not a rewrite. Pick your tool:
- Claude Code — set the base URL via environment variables
- Cursor — override the API endpoint in settings
- OpenAI / Anthropic SDK — one
base_urlargument - Python — sync, streaming and async examples
- Cutting the bill — eight levers ranked by effect
FAQ
How much does GLM-5.3 cost per million tokens?
$0.7 per million input tokens and $2.2 per million output tokens. No official USD list price is published for this model, so it is billed at platform rate.
Is GLM-5.3 cheaper than buying direct?
There is no published official USD price to compare against for this model — it is billed at platform rate, and the reason to route it through a gateway is access and consolidated billing rather than a discount.
What does GLM-5.3 cost per month in practice?
At a 5:1 input-to-output ratio it is about $5.70 a month for light use (5M input tokens) and roughly $570.00 for heavy use (500M input). Agent workloads sit at the high end because every turn resends the context — see what actually drives Claude Code cost.
Which models cost about the same as GLM-5.3?
The comparison table further up this page lists the three closest in blended price. If you are choosing on cost alone, start there and test the top two on your own prompts — price per token is only half the equation, the other half is how many tokens a model needs to finish the task.
How do I start using GLM-5.3?
You need an API key, then point your client at the gateway base URL. It is a one-line change in most SDKs — see the SDK setup guide, or the Claude Code and Cursor guides for those tools.