Model Studio 401s: pair the API key with the right plan, region, and base URL path

TL;DR: On a 401, check three things: the key belongs to the same region and plan as the endpoint, the host matches your region (intl hosts for Singapore/US), and the path matches your client (/compatible-mode/v1 for OpenAI-compatible clients, /api/v1 for the DashScope SDK). For production use the workspace-dedicated domain, not the trial domain, to get proper rate limits.

Fix

1. On a 401, check three things: the key belongs to the same region and plan as the endpoint, the host matches your region (intl hosts for Singapore/US), and the path matches your client (/compatible-mode/v1 for OpenAI-compatible clients, /api/v1 for the DashScope SDK). Expected: this specific failure stops.

2. For production use the workspace-dedicated domain, not the trial domain, to get proper rate limits. Expected: the request routes instead of erroring.

When to use

You hit exactly this: Model Studio 401s: pair the API key with the right plan, region, and base URL path in Model Studio 401s.

When not to use

A different error, or the same symptom in a different tool. This page only covers the failure above.

Compatibility

Model Studio 401s (versions mentioned here: v1).

Documented gotcha from official Alibaba Cloud Model Studio docs: Official docs (Alibaba Cloud Model Studio base URL): documents that an API key must be paired with an endpoint from the same billing plan (a mismatch returns 401), that keys are region-specific, that the legacy DashScope domain is being replaced by workspace-dedicated domains (recommended) with trial domains carrying lower rate limits, and that OpenAI SDKs must use /compatible-mode/v1 while the native DashScope path is /api/v1. The verified fix/answer: On a 401, check three things: the key belongs to the same region and plan as the endpoint, the host matches your region (intl hosts for Singapore/US), and the path matches your client (/compatible-mode/v1 for OpenAI-compatible clients, /api/v1 for the DashScope SDK). For production use the workspace-dedicated domain, not the trial domain, to get proper rate limits. Source: https://help.aliyun.com/en/model-studio/base-url