The unified API for Qwen models
Qwen API is the OpenAI-compatible API for every Qwen model — chat, vision, code, video, and agents.
A single API endpoint gives you access to Qwen 3, Qwen 2.5, QwQ, QVQ, and the full Qwen model family. Drop-in replacement for OpenAI, same SDK, same format — with multimodal understanding, function calling, streaming, and MCP tool use built in.
Models
Qwen 3 · 2.5 · QwQ
Latency
Streaming · Batch
Pricing
Pay-as-you-go
API Format
OpenAI-compatible
Key Feature
Function calling
Overview
What makes Qwen API different from other LLM APIs
The Qwen API Platform is designed for developers who need one consistent interface across the entire Qwen model ecosystem — from lightweight on-device models to the frontier Qwen 3 and reasoning-focused QwQ series.
One API, every Qwen model
Switch between Qwen 3, Qwen 2.5, QwQ, QVQ, and domain-specific models by changing a single parameter — no endpoint migration needed.
Full multimodal support
Images, videos, audio, and documents — all through the same chat completion endpoint. No separate vision API to learn.
Enterprise-grade security
Data encryption in transit and at rest, with optional private endpoint deployment through Alibaba Cloud Model Studio.
Agent & tool use ready
Native function calling, MCP (Model Context Protocol) support, code interpreter, and retrieval-augmented generation out of the box.
Models
The full Qwen model family, accessible through one API
From the flagship Qwen 3 to specialized reasoning, vision, and code models — every Qwen variant is available through the same OpenAI-compatible endpoint. No SDK changes, no new integrations.
Qwen 3
LatestFlagship model with advanced reasoning, multilingual capability, and 235B parameters. Supports tool use, code execution, and long-context tasks.
Qwen 2.5
StableProven workhorse in sizes from 0.5B to 72B. Ideal for production deployments where latency and cost are primary considerations.
QwQ
ReasoningDeep thinking model for complex math, logic, and multi-step reasoning tasks. Excels at chain-of-thought and self-verification.
QVQ & VL
VisionMultimodal vision-language models. QVQ adds visual reasoning; Qwen2.5-VL handles image and video understanding with spatial awareness.
Qwen Coder
Code-specialized model built on Qwen 2.5. Supports code generation, explanation, debugging, and multi-language programming tasks.
Wan (Image & Video)
Text-to-image and text-to-video generation API. Create high-quality visual content with prompt enhancement and style control.
Qwen API turns model capability into a developer primitive.
Most AI platforms lock you into one modality or one model size. Qwen API gives you the whole family — chat, reasoning, vision, code, video — all behind a single OpenAI-compatible interface.
Compatible
OpenAI format
Modalities
Text · Image · Video
Tools
MCP · Functions · RAG
Deep dive
What Is the Qwen API Platform? How OpenAI-Compatible Model Serving Works
A clear explanation of what the Qwen API Platform is, which models it supports, how authentication works, and how to make your first request with curl or the OpenAI SDK.
Quick Start
How to go from zero to your first Qwen API call
The Qwen API uses the exact same format as OpenAI. If you have used the OpenAI SDK, you already know how to use Qwen API. Here is the shortest path to a working request.
Get your API key
Sign up at the Qwen API console and generate your API key. No credit card required for the free tier.
Call the API with curl
Set your API key and send a chat completion request to the OpenAI-compatible endpoint.
-H "Authorization: Bearer $QWEN_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen-3",
"messages": [{"role": "user", "content": "Hello"}]
}'
Use the OpenAI Python SDK
Just change the base URL — your existing OpenAI code works with Qwen API.
client = OpenAI(
api_key="your-key",
base_url="https://api.qwen.ai/v1"
)
response = client.chat.completions.create(
model="qwen-3",
messages=[{"role": "user", "content": "Hello"}])
print(response.choices[0].message.content)
Try advanced features
Stream responses, attach images, use function calling, or enable MCP tool use — all through the same API endpoint.
FAQ
The fastest answers to the questions developers ask first
Start here if you want the endpoint URL, pricing model, model availability, or authentication details without reading the full docs.
Primary sources
Every claim on this page is grounded in the official Qwen documentation, model cards, or linked resources so you can verify the details yourself.