Models / Qwen3 Omni

Qwen3 Omni

Tier 1Apache-2.0

Real-time speech-capable multimodal model with 234ms first-packet latency. Supports voice-to-voice conversations.

qwen/qwen3-omni
Context Window
131K
Max Output
16K
Providers
1
Released
2025-04

Capabilities

ChatVisionAudioStreaming

Pricing by Provider

Prices shown are the official rate published by the model provider, before any platform discounts. Your actual rate may be lower based on your account's discount tier.

ProviderPriceLatency p50Latency p95Status
alibaba$0.43 /M in · $1.66 /M out234ms600ms

Quick Start

Python
import magicrouter

mr = magicrouter.Client(
    provider_keys={"alibaba": "your-api-key"}
)

response = mr.chat(
    "qwen/qwen3-omni",
    "Your prompt here"
)
print(response.choices[0].message.content)
TypeScript
import { MagicRouter } from "magicrouter";

const mr = new MagicRouter({
  providerKeys: { alibaba: "your-api-key" }
});

const response = await mr.chat({
  model: "qwen/qwen3-omni",
  messages: [{ role: "user", content: "Your prompt here" }]
});
console.log(response.choices[0].message.content);
cURL
curl https://dashscope-intl.aliyuncs.com/compatible-mode/v1/chat/completions \
  -H "Authorization: Bearer your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3-omni-flash",
    "messages": [{"role": "user", "content": "Your prompt here"}]
  }'

Use this model

Sign up for free and test Qwen3 Omni in the playground

Get Started