ModelSmarterStart intake
ModelsPlansDocs
Browse documentationTap to expand
Documentation overview

API setup

OpenAI-Compatible Chat APIOpenAI-Compatible Responses APIBatch and Offline API JobsRate Limits and Quota Errors

All sections

Scroll
Product basicsBilling and pricingAPI setupModel inferenceDeveloper toolsOperationsSecurity and complianceTutorials and support

API setup

OpenAI-Compatible Chat API

Most OpenAI-compatible integrations need only three changes: API key, base URL, and model name. The hard part is choosing the correct plan and endpoint.

API setupOfficial source

Who this is for

Developers migrating an existing OpenAI-style application.

Configuration reference

Values to confirm before setup

International / Singapore

https://dashscope-intl.aliyuncs.com/compatible-mode/v1

United States / Virginia

https://dashscope-us.aliyuncs.com/compatible-mode/v1

China / Beijing

https://dashscope.aliyuncs.com/compatible-mode/v1

China / Hong Kong

https://cn-hongkong.dashscope.aliyuncs.com/compatible-mode/v1

Token Plan Team Edition

https://token-plan.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1

Token Plan Anthropic-compatible

https://token-plan.ap-southeast-1.maas.aliyuncs.com/apps/anthropic

Setup flow

Practical steps

  1. 01Install or keep the OpenAI-compatible SDK.
  2. 02Set the API key in an environment variable.
  3. 03Replace the base URL with the correct Model Studio or Token Plan endpoint.
  4. 04Replace the model name with a supported Qwen/model identifier.
  5. 05Run non-streaming first, then enable streaming or tools if needed.
  6. 06Add error logging for 401, model-not-found, rate-limit, and quota errors.

The three values that must match

Most connection failures come from mixing a key, endpoint, and model from different plans or regions. Confirm the API key source first, then choose the matching base URL, then use a model name supported by that route.

  • API key: Model Studio pay-as-you-go, Token Plan, and Coding Plan keys are not interchangeable.
  • Base URL: region-specific endpoints are not universal.
  • Model name: use the exact published model identifier, including suffixes such as qwen3.7-plus or glm-5.1 where required.

Security baseline

Use environment variables or an application secret manager. Do not paste production API keys into public code, screenshots, frontend JavaScript, or customer-facing support tickets.

SDK vs HTTP endpoint

SDK clients usually use a base URL ending in `/compatible-mode/v1`. Raw HTTP chat completions calls append `/chat/completions` to the endpoint.

Examples

Copy the structure, not the secrets

Node.js OpenAI-compatible client

ts

Scroll
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.DASHSCOPE_API_KEY,
  baseURL: "https://dashscope-intl.aliyuncs.com/compatible-mode/v1",
});

const completion = await client.chat.completions.create({
  model: "qwen3.7-plus",
  messages: [{ role: "user", content: "Create a concise onboarding checklist." }],
});

console.log(completion.choices[0]?.message?.content);

Common mistakes

Check these before escalating

  • Do not mix Token Plan keys with pay-as-you-go endpoints.
  • Some models support Responses API while others need Chat Completions.
  • Tool clients may require a different protocol such as Anthropic-compatible.

Related guides

OpenAI-Compatible Responses API

Responses API is useful for newer agent-style clients and models that support the OpenAI Responses wire format. Confirm support before choosing it for a tool.

First Qwen API Call

A first API call needs an activated account, Model Studio API key, region-specific base URL, OpenAI SDK or HTTP client, and a model that is available in the chosen region.

Rate Limits and Quota Errors

Rate limits are calculated by account, model, and aggregate API-key usage. A customer quote should include traffic assumptions and an escalation path.

Codex

Configure Codex with Model Studio or Token Plan credentials, then verify the selected model before handing it to developers.