Is Pony Alpha Free? GLM-5 Pricing Explained
Do not assume the released GLM-5 model is free because Pony Alpha was once available for experimental access. OpenRouter identifies Pony Alpha as an early GLM-5 testing version, while released GLM-5 has a separate listing with current provider prices. You should confirm the rate on the provider you will actually use before sending requests.
Rates and promotions can change. Pricing checked September 23, 2026; verify again before publishing or purchasing.
The quick answer
| Question | Answer |
|---|---|
| Was the original Pony Alpha test offered for free? | It was described as an experimental zero-price listing when users first explored it. Historical terms do not guarantee current access. |
| Is GLM-5 free on OpenRouter? | The current GLM-5 listing shows paid token rates and multiple providers. Check it for today's price and any free variant. |
| Is GLM-5 free at Z.AI? | Z.AI's pricing table lists per-token charges for GLM-5. Check your account for any separate promotion or plan. |
| Do providers charge the same amount? | No. Different providers or plans may show different prices, cache terms, or promotions. |
Why the old “free Pony Alpha” claim is misleading now
Pony Alpha was an early test listing under openrouter/pony-alpha. OpenRouter later linked it to the released model z-ai/glm-5. These are distinct listings. A screenshot of the old test price does not establish today's GLM-5 rate or whether the old endpoint accepts requests. See what happened to Pony Alpha and our OpenRouter access guide.
What does GLM-5 currently cost?
Provider snapshot, September 23, 2026: OpenRouter's GLM-5 page displays a starting promotional rate of $0.60 per million input tokens and $1.92 per million output tokens, while its provider table lists other rates. Z.AI's direct pricing page lists $1 per million input tokens and $3.20 per million output tokens for GLM-5, with a separate cached-input rate. These are two different purchasing routes. Promotions, routing, and rates may change; open OpenRouter pricing and Z.AI pricing before publishing or paying.
How does token billing work?
API providers typically price input tokens (what you send) and output tokens (what the model returns) separately. The total depends on how much text and context you send, how long the response is, any cached-input rules, and how often you retry. Check your provider dashboard for your actual charge.
Example calculation only: At a rate of $0.60 per million input tokens and $1.92 per million output tokens, a request using 10,000 input tokens and 2,000 output tokens would cost roughly $0.00984 before any other applicable charges: (10,000 / 1,000,000 × $0.60) + (2,000 / 1,000,000 × $1.92). This is an illustration using one listed rate, not a quote for every provider.
Is a free alternative available?
Look for an explicitly marked free model variant in the current OpenRouter catalog or a current provider promotion. Do not assume a free listing has the same model, limits, privacy terms, or reliability as released GLM-5. If your main goal is a working replacement, start by testing the released model with a small task and recording its actual cost. See Pony Alpha alternatives if your budget or use case calls for a different model.
Avoid surprise costs
Check the live provider and price directly before creating a key or sending a request.
Set a spending limit or alerts if your provider offers them.
Start with short nonsensitive test prompts and check actual token usage.
Limit unnecessary context and repeated calls, while keeping enough information for the model to complete the task.
Record the whole task cost, including retries, rather than comparing input-token prices alone.
FAQs
Is the old Pony Alpha model still free to use?
The old listing's historical price does not establish its current availability or cost. Check the original OpenRouter listing and your account; use the released GLM-5 listing for a new integration.
Does a ChatGPT subscription include GLM-5?
No claim should be inferred from a ChatGPT subscription about access to an OpenRouter or Z.AI model. Compare actual products and providers in Pony Alpha vs ChatGPT.
Which GLM-5 price should I quote?
Name the provider, pricing date, input and output rates, and any promotion. The OpenRouter and Z.AI figures above are not interchangeable.
How do I try the model before spending much?
Follow our GLM-5 setup guide and make a small test request, then inspect the charge.
Return to the Pony Alpha homepage.