Qwen3-omni-flash model input URL audio error
# helpv
veth0119
09/30/2025, 12:36 PMThe audio file can be previewed correctly in the playground, but the request error occurs.
Copy code
client AliQwen3Omni {
provider openai-generic
options {
model "qwen3-omni-flash"
base_url "https://dashscope.aliyuncs.com/compatible-mode/v1"
api_key env.QWEN_API_KEY
}
}veth0119
09/30/2025, 12:48 PMOfficial curl example
Copy code
curl -X POST https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions \
-H "Authorization: Bearer $DASHSCOPE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen-omni-turbo",
"messages": [
{
"role": "user",
"content": [
{
"type": "input_audio",
"input_audio": {
"data": "https://help-static-aliyun-doc.aliyuncs.com/file-manage-files/zh-CN/20250211/tixcef/cherry.wav",
"format": "wav"
}
},
{
"type": "text",
"text": "What is this audio saying?"
}
]
}
],
"stream":true,
"stream_options":{
"include_usage":true
},
"modalities":["text","audio"],
"audio":{"voice":"Cherry","format":"wav"}
}'veth0119
09/30/2025, 12:50 PMThere are two differences:
1. The baml playground does not pass the URL, but passes the base64 data
2. The format of the baml playground is x-wav, not wav
v
Vaibhav (BAML 🐑)
09/30/2025, 1:41 PMYou might need to add modalities and audio to the options definition
v
veth0119
10/01/2025, 1:32 AM@Vaibhav (BAML 🐑)The actual test results show that there is no need to pass options. I think the problem still lies in the baml URL processing.
https://cdn.discordapp.com/attachments/1422562712172630189/1422757928414482542/image.png?ex=68ddd613&is=68dc8493&hm=d8a032001fc21cff335f91a3c9b343b0d2758b4e2293e9ea8ae03ab86e688e67&
veth0119
10/01/2025, 11:49 PM@Vaibhav (BAML 🐑) any suggestion？
v
Vaibhav (BAML 🐑)
10/03/2025, 12:10 PMSorry for the delay here. It looks like it needs non base64 data which openai doesn’t in their spec.
We can patch this!