Short answer: GLM-4.7 is not free through Z.ai’s standard API—but GLM-4.7-Flash is listed as free
Z.ai’s pricing table lists GLM-4.7 at $0.60 per million input tokens and $2.20 per million output tokens. It lists GLM-4.7-Flash as free. These are separate models: the free listing for Flash does not make standard GLM-4.7 API calls free. Prices are in USD. (docs.z.ai)
Z.ai’s model overview describes GLM-4.7-Flash as “Completely Free” and lists the GLM-4.7 series with a 200K-token context length and 128K maximum output. The pricing page does not specify a free-tier quota, rate limit, service-level commitment, or other production terms for Flash, so do not assume “free” means unlimited or guaranteed production capacity. (docs.z.ai)
Z.ai model |
Input price | Output price | What the published pricing indicates |
|---|---|---|---|
| GLM-4.7 | $0.60 / 1M tokens | $2.20 / 1M tokens | Paid API usage |
| GLM-4.7-Flash | Free | Free | Listed as free; quota and service terms are not stated in the pricing table |
The figures above are from Z.ai’s published API pricing. They do not establish prices or access terms at other providers. (docs.z.ai)
What does “free GLM-4.7” mean?
For a budget decision, distinguish the exact model and access route:
- GLM-4.7 through
Z.ai’s standard API: paid at the listed input and output rates. - GLM-4.7-Flash through
Z.ai’s API: listed as free. Check your account’s current availability and limits before relying on it for an application, because the cited pricing table does not publish those terms. - GLM Coding Plan: a separate subscription, not a free GLM-4.7 API allowance.
Z.aicurrently advertises plans starting at $18 per month, and its plan documentation says requests for GLM-4.7 are automatically routed to GLM-5.3-Flash. That routing means a Coding Plan request labelled GLM-4.7 should not be treated as access to the same model as standard GLM-4.7 API usage. (docs.z.ai) - Model weights and self-hosting:
Z.ai’s official GLM-4.7 model card labels the model’s license as MIT. The MIT license grants use free of charge and permits commercial use, subject to conditions including retaining the copyright and license notices. Check the license accompanying the exact artifact you download. This is a no-fee model license, not free hosting: hardware or cloud infrastructure still costs money, and hosted API calls are billed separately. (opensource.org)
In short, if you mean standard GLM-4.7 API calls, budget for paid tokens. If you mean a free GLM-series API option, Z.ai lists GLM-4.7-Flash as free—but confirm the limits and production terms for your account.
GLM-4.7 API cost example
Z.ai lists prices per million tokens. For an illustrative request volume of 1 million input tokens and 250,000 output tokens, using the published standard GLM-4.7 rates and assuming no cached-input discount:
- Input: 1 × $0.60 = $0.60
- Output: 0.25 × $2.20 = $0.55
- Estimated total: $1.15
This is a worked estimate from the published rates, not a measured bill. Actual usage depends on token counts and applicable account pricing. Z.ai’s table also lists cached input at $0.11 per million tokens and cached-input storage as limited-time free; those are separate pricing entries, so check how your request qualifies before applying them to an estimate. (docs.z.ai)
A simple planning formula is:
Estimated API cost = (input tokens ÷ 1,000,000 × $0.60) + (output tokens ÷ 1,000,000 × $2.20)
For example, for 10 million input tokens and 2 million output tokens, the uncached estimate is $6.00 + $4.40 = $10.40. This excludes any other charges and assumes the listed rates apply to your account and endpoint.
Is a free option suitable for production?
Free token pricing alone does not establish production suitability. Before choosing GLM-4.7-Flash or budgeting for paid GLM-4.7, check the specific endpoint and model available to your account, your expected throughput and usage limits, and any service terms that matter to your application. Z.ai documents GLM-4.7 capabilities including coding, multi-step reasoning, streaming, function calling, and structured output; treat those as product documentation, not a guarantee that it will meet your workload’s quality or reliability requirements. (docs.z.ai)
A practical evaluation:
- Select the exact route. Decide between standard GLM-4.7 API, GLM-4.7-Flash, a subscription plan, or self-hosting. Do not compare prices across these as if they were the same offer.
- Estimate token volume. Forecast monthly input and output tokens separately, then use the matching rates. For Flash, verify applicable quotas and limits rather than assuming unlimited free calls.
- Test representative tasks. Run a small sample of the prompts, tool calls, or coding tasks your application actually needs. No benchmark result can determine fit for every workload.
- Budget for the paid route if limits matter. If your workload needs standard GLM-4.7, use its paid rates in the forecast; do not assume a free Flash listing or Coding Plan subscription covers standard API calls.
Bottom line
GLM-4.7 API: paid. GLM-4.7-Flash: listed as free by Z.ai. For standard GLM-4.7, the published rates are $0.60 per million input tokens and $2.20 per million output tokens. The official GLM-4.7 model card labels its model license MIT; that does not make hardware or hosted API usage free. Treat the free Flash option as a separate model, and confirm its applicable limits and production terms before depending on it.
Sources
Z.AI Developer Documentation — Pricing
Z.AI Developer Documentation — GLM-4.7 Overview
Z.AI Developer Documentation — GLM Coding Plan Overview
Z.AI Developer Documentation — Quick Start
zai-org/GLM-4.7 — Hugging Face
The MIT License — Open Source Initiative
zai-org/GLM-4.5 — Official GLM-4.7, GLM-4.6, and GLM-4.5 repository