Is GLM-4.6 free?
Not for regular token-billed API use on Z.ai: its published pricing lists GLM-4.6 at $0.60 per million input tokens and $2.20 per million output tokens. Z.ai’s pricing page says its prices are in USD. It labels cached-input storage “Limited-time Free,” but that applies to cache storage—not to GLM-4.6 inference or all API calls. (docs.z.ai)
I did not verify an ongoing free production API allowance, free credits, or a trial for GLM-4.6. That does not prove no provider ever offers a promotion; check the terms for the specific provider and access channel before relying on one. A free chat interface, promotional credits, or locally run model weights would not, by themselves, mean that a hosted production API is free.
GLM-4.6 API prices and an example budget
Z.ai’s listed token rates are:
| Usage | Z.ai listed price |
|---|---|
| Input tokens | $0.60 per 1 million tokens |
| Cached input tokens | $0.11 per 1 million tokens |
| Cached-input storage | Limited-time free, according to the pricing table |
| Output tokens | $2.20 per 1 million tokens |
These are Z.ai’s published rates, in USD. The “free” entry is specific to cached-input storage; it does not make ordinary input or generated output free. (docs.z.ai)
Illustrative calculation, not a measured workload: suppose an application uses 1 million uncached input tokens and generates 250,000 output tokens in a billing period. At the listed rates, the token charges would be:
- Input: 1 × $0.60 = $0.60
- Output: 0.25 × $2.20 = $0.55
- Total: $1.15
This estimate assumes those published rates apply to your account and usage. It excludes any other services or charges, and does not apply the cached-input rate. Actual usage depends on token counts and applicable account terms.
Provider and channel matter. For example, Together AI’s GLM-4.6 page also displays input and output rates, but says GLM-4.6 is not available on its Serverless API; the page does not make clear which deployment its displayed rates apply to. Don’t treat those rates as confirmed Serverless pricing—or as proof of access on another provider. (pricepertoken.com)
Is it available for production API use?
Z.ai’s GLM-4.6 documentation gives the model identifier as glm-4.6 and describes a 200K-token context window and a maximum output of 128K tokens. Z.ai’s launch post says the model is available through its API platform. Those details establish documented model and API support, but they do not establish that GLM-4.6 is free. (docs.z.ai)
For a production decision, confirm the current price and account terms with the provider you plan to use, then estimate cost using your expected input and output token volumes. If you’re considering another host, verify its model identifier, availability, rates, and any limits separately. A model’s published context capacity does not establish that it will meet your application’s quality, latency, or reliability needs; test it against your workload before committing.
Sources
Z.AI Developer Documentation: Pricing
Z.AI Developer Documentation: GLM-4.6
Z.ai: GLM-4.6 release announcement