Alibaba's "Tongyi-MAI/Z-Image-Turbo" model page provides API-Inference examples that can be directly replicated to facilitate the integration of text-to-text images into automated drawing, content production, and prototyping processes. The platform rules mention that each registered user can currently make a total of 2,000 API-Inference free calls per day, so it is regarded by many as a "large volume and low threshold" mapping solution.
At the same time, this free quota is not equivalent to "a fixed free 2,000 images for a certain model". The platform may also set an independent upper limit for a single model, and reserve room to adjust the quota and concurrency limit at any time. Accounts usually need to complete prerequisites such as real-name authentication. When using batch builds, you should still pay attention to the quota consumption caused by failed retries, the compliance requirements of the generated content, and the differences in the requirements for stability and SLA in the production environment.
FAQs
Q: Can Z-Image-Turbo really generate 2000 images per day for free?
A: The platform describes a total of 2,000 API-Inference calls per day, which may not be equivalent to a single model or "number of sheets" guarantee, and may also be superimposed on the upper limit of a single model and the consumption of failed retries.
Q: Which models are eligible for the free 2000 credits of ModelScope?
A: Quotas are an overall limit at the API-Inference level, covering the available inference models, but some models may have additional individual limits.
Q: Where is the API code copied from Z-Image-Turbo?
A: You can directly copy the sample code and the required parameters in the Inference/Call area of the Z-Image-Turbo model page.
Q: What conditions do I need to meet for my account to use the free credit?
A: The platform rules prompt that you need to complete account requirements such as real-name authentication, and you may need to bind relevant account information to call normally.
Q: What are the common risks of using free inference to do batch mapping?
A: Common risks include quota changes with policies, queuing due to concurrency limits, additional consumption for failed retries, and unclear boundaries between compliance audits and commercial authorization.
Official address: https://modelscope.cn/models/Tongyi-MAI/Z-Image-Turbo
