Notice on New Models hy4 and deepseek-v4.1-flash-1017 in the Mini Program Growth Plan

The AI Resource Pack gifted by the Mini Program Growth Plan now covers two more models: hy4 and deepseek-v4.1-flash-1017. Together with the existing hy4-preview, hy3, and deepseek-v4.1-flash, five text models can currently consume the gifted quota.

• No enablement required: none of the five models needs to be enabled in the CloudBase console — call them directly.
• Automatic deduction: when called from an allowed source (mini program, CloudBase server-side, CloudBase console), Token usage is deducted from the gifted AI Resource Pack automatically, with no extra configuration.
• Two code changes required: ① set the createModel argument to cloudbase; ② set the model parameter to one of the values in the table below.

Covered Models

model Parameter	Context	Image	Video	File	Deep Thinking	Tool Calling	Structured Output	Caching
hy4	1M	✅	❌	—	✅	✅	✅	✅
hy4-preview	1M	❌	❌	—	✅	✅	✅	✅
hy3	256K	❌	❌	—	✅	✅	✅	✅
deepseek-v4.1-flash	1M	✅	—	—	✅	✅	✅	✅
deepseek-v4.1-flash-1017	1M	✅	—	—	✅	✅	✅	✅

✅ Supported, ❌ Not supported, — Not stated by the vendor for that model (not necessarily unsupported).

hy4 is the official release of the Hunyuan model: compared with hy4-preview, it strengthens Agent and long-horizon task capabilities and introduces image understanding for the first time. deepseek-v4.1-flash-1017 is the 1017 version snapshot of deepseek-v4.1-flash.

Code Example

const model = ai.createModel("cloudbase");
const res = await model.invoke({
  model: "hy4",
  messages: [{ role: "user", content: "Hello" }],
});

For usage details, see the Mini Program Growth Plan User Guide. For the allowed scope of the quota, see the Special Governance Announcement.

---

Thank you for your support and understanding! If you have any questions, please visit the CloudBase Community. We will address your inquiries as soon as possible!