Alibaba Cloud Model Studio:Error messages

400-InvalidParameter

parameter.enable_thinking must be set to false for non-streaming calls/parameter.enable_thinking only support stream call

Cause: A thinking model was called using a non-streaming output method.

Solution: Set the enable_thinking parameter to false, or use the streaming method.

The thinking_budget parameter must be a positive integer and not greater than xxx

Cause: The value of the thinking_budget parameter is outside the valid range.

Solution: For the maximum chain-of-thought length for the model, see Models. Specify a value that is greater than 0 and does not exceed this length limit.

This model only support stream mode, please enable the stream parameter to access the model.

Cause: The model supports only streaming output, but streaming output was not enabled in the call.

Solution: Call the model using the streaming output method.

This model does not support enable_search.

Cause: The current model does not support the web search feature, but the enable_search parameter was set to true.

Solution: Call a model that supports the web search feature.

暂时不支持当前设置的语种！(The currently set language is not supported!)

Cause: When you use the Qwen-MT model, the format of the source_lang or target_lang parameter is incorrect, or the language is not in the list of supported languages.

Solution: Enter the English name or language code in the correct format.

The incremental_output parameter must be "true" when enable_thinking is true

Cause: When thinking mode is enabled for a model, it only supports incremental streaming output. However, the incremental_output parameter was not set to true.

Solution: Set the incremental_output parameter to true and try again. The API returns incremental content.

The incremental_output parameter of this model cannot be set to False.

Cause: The model supports only incremental streaming output, but the incremental_output parameter was not set to true.

Solution: Set the incremental_output parameter to true and try again. The API returns incremental content.

Range of input length should be [1, xxx]

Cause: The input length exceeds the model's limit.

Solution:

• If you call the model through code, make sure that the number of tokens in the messages array is within the model's maximum input token limit.

• When you have a continuous conversation in a chat client, such as Chatbox, or in the Model Studio console, each request includes the conversation history. This can cause the request to exceed the model's limit. If the limit is exceeded, you must start a new conversation.

Range of max_tokens should be [1, xxx]

Cause: The value of the max_tokens parameter is not in the range of [1, maximum output tokens for the model].

Solution: Set the max_tokens parameter based on the model's maximum output token limit.

Temperature should be in [0.0, 2.0)/'temperature' must be Float

Cause: The value of the temperature parameter is not in the range of [0.0, 2.0).

Solution: Set the temperature parameter to a number greater than or equal to 0 and less than 2.

Range of top_p should be (0.0, 1.0]/'top_p' must be Float

Cause: The value of the top_p parameter is not in the range of (0.0, 1.0].

Solution: Set the top_p parameter to a number greater than 0 and less than or equal to 1.

Parameter top_k be greater than or equal to 0

Cause: The top_k parameter is set to a number less than 0.

Solution: Set the top_k parameter to a number greater than or equal to 0.

Repetition_penalty should be greater than 0.0

Cause: The repetition_penalty parameter is set to a number less than or equal to 0.

Solution: Set the repetition_penalty parameter to a number greater than 0.

Presence_penalty should be in [-2.0, 2.0]

Cause: The value of the presence_penalty parameter is not in the range of [-2.0, 2.0].

Solution: Set the presence_penalty parameter to a value in the range of [-2.0, 2.0].

Range of n should be [1, 4]

Cause: The value of the n parameter is not in the range of [1, 4].

Solution: Set the n parameter to a value in the range of [1, 4].

Range of seed should be [0, 9223372036854775807]

Cause: When you use the DashScope protocol, the value of the seed parameter is not in the range of [0, 9223372036854775807].

Solution: Set the seed parameter to a value in the range of [0, 9223372036854775807].

Request method 'GET' is not supported.

Cause: The current API does not support the GET request method.

Solution: Refer to the API reference and send the request again using a supported request method, such as POST.

Messages with role "tool" must be a response to a preceding message with "tool_calls"

Cause: An assistant message was not added to the messages array during tool calling.

Solution: Add the assistant message from the model's first-turn response to the messages array before you add the tool message.

Required body invalid, please check the request body format.

Cause: The request body format does not meet the API requirements.

Solution: Check the request body to make sure that it is a standard JSON string. Common issues include extra commas (,) or unclosed parentheses. You can use an large language model to help fix the request body format.

input content must be a string.

Cause: Text-only models do not support setting the content in messages to a non-string type.

Solution: Do not set the content to an array type, such as [{"type": "text","text": "Who are you?"}].

The content field is a required field.

Cause: The content parameter was not specified in the request, for example, {"role": "user"}.

Solution: Specify the content parameter. For example, {"role": "user","content": "Who are you"}.

current user api does not support http call.

Cause: The current model does not support non-streaming output.

Solution: Use streaming output.

Either \"prompt\" or \"messages\" must exist and cannot both be none

Cause: When you call a large language model, neither the messages parameter nor the prompt parameter (deprecated) was specified. If this error occurs after you specify the messages parameter, the format might be incorrect. For example, when you use DashScope-HTTP, messages must be placed in the input object, not at the same level as the model parameter.

Solution: Specify the messages parameter. If you have specified the parameter but the error persists, see the Qwen API reference to check whether its position is correct.

'messages' must contain the word 'json' in some form, to use 'response_format' of type 'json_object'.

Cause: When you use structured output, the prompt does not contain the json keyword.

Solution: Add json (case-insensitive) to the prompt, such as "Output in json format."

Json mode response is not supported when enable_thinking is true

Cause: The model's thinking mode was enabled when you use structured output.

Solution: When you use structured output, set enable_thinking to false to disable thinking mode. For more information, see How do models in thinking mode produce structured output?

Tool names are not allowed to be [search]

Cause: The tool name cannot be set to search.

Solution: Set the tool name to a value other than search.

Unknown format of response_format, response_format should be a dict, includes 'type' and an optional key 'json_schema'. The response_format type from user is xxx.

Cause: The specified response_format parameter is invalid.

Solution: To use the structured output feature, set the response_format parameter to {"type": "json_object"}.

The value of the enable_thinking parameter is restricted to True.

Cause: For some models, such as qwen3-235b-a22b-thinking-2507, the enable_thinking parameter cannot be set to false.

Solution:

• If you call the model from a third-party tool, such as Cherry Studio, turn on the thinking switch in the input box.

• If you call the model from code, set enable_thinking to true.

'audio' output only support with stream=true

Cause: The Qwen-Omni model only supports streaming output, but you did not use it.

Solution: Set the stream parameter to true.

tool_choice is one of the strings that should be ["none", "auto"]

Cause: The tool_choice parameter specified for a function calling request is incorrect.

Solution: Specify "auto" (the model chooses the tool) or "none" (forces the model not to use a tool).

Model not found

Cause: The specified model parameter does not exist.

Solution: The case of the model parameter may be incorrect, or Model Studio may not provide the model that you want to call. You can check the model name in Models to verify that the model you entered is correct.

Make sure to use model names from Model Studio instead of open source communities. For example, use qwen3-235b-a22b-instruct-2507 instead of Qwen/Qwen3-235B-A22B-Instruct-2507.

The result_format parameter must be \"message\" when enable_thinking is true

Cause: When you call a thinking mode model, the result_format parameter was not set to "message".

Solution: Set the result_format parameter to "message".

The audio is empty

Cause: The input audio is too short, which results in an insufficient number of sampling points.

Solution: Increase the duration of the audio.

File parsing in progress, please try again later.

Cause: When you use the Qwen-Long model, the file has not finished parsing.

Solution: Wait for the file to finish parsing and then try to call the API again.

The "stop" parameter must be of type "str", "list[str]", "list[int]", or "list[list[int]]", and all elements within the list must be of the same type.

Cause: The stop parameter does not conform to the str, list[str], list[int], or list[list[int]] format.

Solution: Refer to the Qwen API reference and set the stop parameter to the correct format.

Value error, batch size is invalid, it should not be larger than xxx.

Cause: When you call an embedding model, the number of texts exceeds the model's limit.

Solution: Refer to the batch size for the model in Embedding. Then, control the number of input texts.

[] is too short

Cause: The input messages array is empty.

Solution: Add a message before you send the request.

The tool call is not supported.

Cause: The model does not support the tools parameter.

Solution: Switch to a Qwen or DeepSeek model that supports tool calling.

Required parameter(xxx) missing or invalid, please check the request parameters.

Cause: The API call parameters are invalid.

Solution: Check the request parameters to make sure that all required parameters are provided and in the correct format.

input must contain file_urls

Cause: When you use audio file recognition with speech recognition (Paraformer), the file_urls request parameter was not assigned a value.

Solution: Include the file_urls parameter in the request and assign it a value.

The provided URL does not appear to be valid. Ensure it is correctly formatted.

Cause: When you use a visual understanding, omni-modal, or audio understanding model, the URL or local path of the input data is invalid or does not meet the requirements.

Solution:

• Input URL: The URL must start with http://, https://, or data:. If the URL starts with data:, "base64" must be included before the Base64-encoded data.

• Input local path: The path must start with file://.

• Input temporary URL:

  • For HTTP calls, make sure that the parameter X-DashScope-OssResourceResolve: enable is added to the request header.

  • For SDK calls, only DashScope SDK calls are supported. Do not use the OpenAI SDK.

Input should be a valid dictionary or instance of GPT3Message

Cause: The format of the messages field does not meet the requirements. For example, the number of parentheses does not match or required key-value pairs are missing.

Solution: Check that the JSON structure of the messages field is correct.

Value error, contents is neither str nor list of str.: input.contents

Cause: When you use an embedding model, the input is not a string or an array of strings.

Solution: Change the input format to a string or a list of strings.

Failed to download multimodal content. / Download the media resource timed out during the data inspection process. / Unable to download the media resource during the data inspection process.

Cause: Failed to download the image, video, or audio, or the download timed out.

• Connectivity issues: You are using an internal network address from Object Storage Service.

• Network latency: Cross-region access causes timeouts.

• Service instability: The source storage service responds slowly or is unreachable.

Solution:

• Change the storage service

  Use a storage service that is in the same region as the model service. Use Object Storage Service to generate public URLs. Do not use private network addresses.

• Adjust the transfer method

  If you cannot transfer the file using a public URL, switch to a recommended method as shown in the table below. For more information, see Use local files (Base64 encoding or file path).

  Type	Specifications	DashScope SDK (Python, Java)	OpenAI-compatible / DashScope HTTP
  Image	Greater than 7 MB and less than 10 MB	Pass the local path	Only public URLs are supported. Use Object Storage Service.
  	Less than 7 MB	Pass the local path	Base64 encoding
  Video	Greater than 100 MB	Only public URLs are supported. Use Object Storage Service.	Only public URLs are supported. Use Object Storage Service.
  	Greater than 7 MB and less than 100 MB	Pass the local path	Only public URLs are supported. Use Object Storage Service.
  	Less than 7 MB	Pass the local path	Base64 encoding
  Audio	Greater than 10 MB	Only public URLs are supported. Use Object Storage Service.	Only public URLs are supported. Use Object Storage Service.
  	Greater than 7 MB and less than 10 MB	Pass the local path	Only public URLs are supported. Use Object Storage Service.
  	Less than 7 MB	Pass the local path	Base64 encoding

  Base64 encoding increases the data size. Therefore, the original file must be smaller than 7 MB.

  Using Base64 encoding or a local file path can prevent server-side download timeouts and improve stability.

url error, please check url!

Reason: Possible causes:

• Outdated DashScope SDK version: Older SDK versions cannot detect the correct server-side address when calling image or video generation models.

• Mismatched interface and model: For example, using a multimodal interface such as MultiModalConversation.call() to call a plain text model such as qwen3-max, or providing an incorrect URL when sending a POST request directly.

Solution: Select the correct interface for the model.

If a SDK call to the image/video generation model returns an error, upgrade the SDK to the latest version.

The following interfaces are prone to errors:

• Multimodal models (such as qwen3-v-plus, qwen-vl-max): MultiModalConversation.call() and POST https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation.

  If using the spring-ai-alibaba framework, confirm whether the multimodal parameter withMultiModel is set.

  Reference: Visual understanding.

• Text-only models (such as qwen3-max, qwen-plus, deepseek-v3.2): Generation.call() and POST https://dashscope.aliyuncs.com/api/v1/services/aigc/text-generation/generation.

  Reference: Text generation.

• CosyVoice voice cloning: The model and target_model parameters: set the model parameter to voice-enrollment, and the target_model parameter to a specific CosyVoice model.

  Reference: CosyVoice voice cloning.

Check the relevant API reference and use the corresponding interface.

check input data style.

Cause: An input parameter does not meet the requirements.

Solution: Check and correct the input parameters.

An error during model pre-process.

Cause: The content field was passed in an incorrect format.

Solution:

• If you are calling from code, do not set the content to an array type, such as [{"type": "text", "text": "Who are you?"}].

The image size is not supported for the data inspection.

Cause:

• The dimensions (length and width) of the image passed to the Qwen-VL model do not meet the model's requirements.

• The output image size exceeds the limit (10 MB).

Solution:

• The image dimensions must meet the following requirements:

  • The width and height of the image must both be at least 10 pixels.

  • The aspect ratio must not exceed 200:1 or 1:200.

• Adjust the parameters for generating the image.

Required parameter(data_sources) missing or invalid, please check the request parameters.

Cause: This error is returned when you call the SubmitIndexJob API because the required parameter SourceType was not specified when you call the CreateIndex API.

Solution: When you create a knowledge base from a given document, this parameter must be set to DATA_CENTER_FILE. When you create a knowledge base from a given category, this parameter must be set to DATA_CENTER_CATEGORY. For more information, see the CreateIndex documentation.

Wrong Content-Type of multimodal url

Cause: The Content-Type field in the response header of the URL request is incorrect.

The Qwen-VL model supports the following Content-Types: image/bmp, image/icns, image/x-icon, image/jpeg, image/jp2, image/png, image/sgi, image/tiff, and image/webp. For more information, see Images supported by the Qwen-VL model.

Solution:

Field required: image_url

Cause: Missing input parameter image_url.

Solution: Refer to Emoji video generation and specify image_url.

Field required: driven_id

Cause: Missing input parameter driven_id.

Solution: Refer to Emoji video generation and specify driven_id.

Invalid ext_bbox

Cause: The ext_bbox input parameter is invalid.

Solution: Refer to Emoji video generation and specify a valid ext_bbox.

Driven not exist: driven_id

Cause:The driven_id input parameter does not exist.

Solution: Refer to Emoji video generation and specify a valid driven_id.

400-invalid_request_error-invalid_value

-1 is lesser than the minimum of 0 - 'seed'/'seed' must be Integer

Cause: When you use the OpenAI compatible protocol, the value of the seed parameter is not in the range of [0, 2³¹-1].

Solution: Set the seed parameter to a value in the range of [0, 2³¹-1].

400-invalid_request_error

you must provide a model parameter.

Cause: The model parameter was not provided in the request.

Solution: Add the model parameter to the request.

400-InvalidParameter.NotSupportEnableThinking

The model xxx does not support enable_thinking.

Cause: The current model does not support setting the enable_thinking parameter.

Solution: Remove the enable_thinking parameter from the request, or use a model that supports thinking mode.

400-invalid_value

The requested voice 'xxx' is not supported.

Cause: When using Qwen-TTS for real-time synthesis, the voice you selected is produced by a different model.

Solution: Make sure the request parameter target_model for voice cloning is the same as the request parameter model for speech systhesis.

400-Arrearage

Access denied, please make sure your account is in good standing.

Cause: Access is denied because your Alibaba Cloud account has an overdue payment.

Solution: Go to the Expenses and Costs page to check for overdue payments.

• No overdue payments: Verify that the API key belongs to your current account.

• Overdue payments: Add funds to your account. Wait a few moments for your account balance to update, and then retry.

400-DataInspectionFailed/data_inspection_failed

Input or output data may contain inappropriate content. / Input data may contain inappropriate content. / Output data may contain inappropriate content.

Cause: The input or output contains content that is suspected to be sensitive and was blocked by Content Moderation.

Solution: Modify the input content and try again.

Input xxx data may contain inappropriate content.

Cause: The input data, such as the prompt or image, may contain sensitive content. Solution: Perform a content compliance check. Modify the input and try again.

400-APIConnectionError

Connection error.

Cause: An on-premises network issue, usually because a proxy is enabled.

Solution: Disable or restart the proxy.

400-InvalidFile.DownloadFailed

The audio file cannot be downloaded.

Cause: When you use audio file recognition with speech recognition (Paraformer), the file to be recognized failed to download.

Solution: Check whether the audio file URL can be accessed from the public network.

400-InvalidFile.AudioLengthError

Audio length must be between 1s and 300s.

Cause: The audio length does not meet the requirement.

Solution: Make sure that the audio duration is in the range of [1, 300] seconds.

400-InvalidFile.NoHuman

The input image has no human body. Please upload other image with single person.

Cause: There is no person in the input image or no face was detected.

Solution: Upload a photo of a single person.

400-InvalidFile.BodyProportion

The proportion of the detected person in the picture is too large or too small, please upload other image.

Cause: The proportion of the person in the uploaded image does not meet the requirement.

Solution: Upload an image where the person's proportion meets the requirement.

400-InvalidFile.FacePose

The pose of the detected face is invalid, please upload other image with whole face and expected orientation.

Cause: The facial pose of the person in the uploaded image does not meet the requirement. The face must be visible and the head orientation must not be severely skewed.

Solution: Upload an image that meets the requirements.

The pose of the detected face is invalid, please upload other image with the expected oriention.

Cause: The facial pose of the person in the uploaded image does not meet the requirement. The head orientation must not be skewed.

Solution: Make sure that the person's face in the image is not skewed.

400-InvalidFile.Resolution

The image resolution is invalid, please make sure that the largest length of image is smaller than 7000, and the smallest length of image is larger than 400.

Cause: The uploaded image size does not meet the requirement.

Solution: The resolution of the uploaded image must not be higher than 7,000 × 7,000 and must not be lower than 400 × 400.

The image resolution is invalid, please make sure that the largest length of image is smaller than xxx, and the smallest length of image is larger than yyy.

Cause: The uploaded image size does not meet the requirement.

Solution: The resolution of the uploaded image must not be higher than xxx × xxx and must not be lower than yyy × yyy.

The image resolution is invalid, please make sure that the aspect ratio is smaller than xxx, and largest length of image is smaller than yyy.

Cause: The uploaded image size does not meet the requirement.

Solution: The aspect ratio of the uploaded image must be less than xxx, and the resolution must not be higher than yyy × yyy.

Invalid video resolution. The height or width of video must be xxx ~ yyy.

Cause: The video resolution does not meet the requirement.

Solution: The length of the video's sides must be between xxx and yyy.

400-InvalidFile.FPS

Invalid video FPS. The video FPS must be 15 ~ 60.

Cause: The video frame rate does not meet the requirement.

Solution: The video frame rate must be between 15 and 60 fps.

400-InvalidFile.Value

The value of the image is invalid, please upload other clearer image.

Cause: The uploaded image is too dark and does not meet the requirement.

Solution: Make sure that the person's face in the image is clear.

400-InvalidFile.FrontBody

The pose of the detected person is invalid, please upload other image with the front view.

Cause: The person in the uploaded image has their back to the camera, which does not meet the requirement.

Solution: Make sure that the person in the image is facing the camera.

400-InvalidFile.FullFace

The pose of the detected face is invalid, please upload other image with whole face.

Cause: The facial pose of the person in the uploaded image does not meet the requirement. The face must be visible.

Solution: Make sure that the person's entire face is visible and not obstructed in the image.

400-InvalidFile.FaceNotMatch

There are no matched face in the video with the provided reference image.

Cause: The face in the reference image does not match the face in the video.

Solution: See VideoRetalk video generation.

400-InvalidFile.Content

The first frame of input video has no human body. Please choose another clip.

Cause: The first frame of the video must contain a person.

Solution: Select a video clip that contains a human body.

The human is too small in the first frame of input video. Please choose another clip.

Cause: The person in the first frame of the video is too small.

Solution: Select a video where the person in the first frame is larger.

The human is not clear in the first frame of input video. Please choose another clip.

Cause: The person in the first frame of the video is not clear.

Solution: Select a video where the person in the first frame is clear.

The input image has no human body or multi human bodies. Please upload other image with single person.

Cause: The input image has no person or multiple people.

Solution: Upload a photo of a single person.

The input image has no human body or has unclear human body. Please upload other image.

Cause: The human body in the input image is incomplete or there is no human body.

Solution: Upload an image that contains a complete and clear human body.

The input image has multi human bodies. Please upload other image with single person.

Cause: The input image has multiple people.

Solution: Upload a photo of a single person.

400-InvalidFile.FullBody

The human is not fullbody in the first frame of input video. Please choose another clip.

Cause: The person in the first frame of the video is not a full body shot.

Solution: The person's entire body must be visible.

The pose of the detected person is invalid, please upload other image with whole body, or change the ratio parameter to 1:1.

Cause: The pose of the person in the uploaded image does not meet the requirement.

Solution: Upload an image that meets the requirements. For a head shot, the entire head must be visible. For a half-body shot, the body from the hips up must be fully visible. Or, change the aspect ratio to 1:1.

400-InvalidFile.BodyPose

The pose of the detected person is invalid, please upload other image with whole body and expected orientation.

Cause: The action of the single person does not meet the requirement.

Solution: Upload an image that meets the requirements. The shoulders and ankles must be visible, the person must not have their back to the camera or be sitting, and the person's orientation must not be severely skewed.

400-InvalidFile.Size

Invalid file size. The video file size must be less than 200MB, and the audio file size must be less than 15MB.

Cause: The file size does not meet the requirement.

Solution: The video file must be smaller than 200 MB, and the audio file must be smaller than 15 MB.

Invalid file size, The image file size must be smaller than 5MB.

Cause: The file size does not meet the requirement.

Solution: The image file must be smaller than 5 MB.

Invalid file size. The video/audio/image file size must be less than xxxMB.

Cause: The file size does not meet the requirement.

Solution: The video, audio, or image file must be smaller than the specified number of megabytes.

400-InvalidFile.Duration

Invalid file duration. The file duration must be xxx s ~ yyy s.

Cause: The file duration does not meet the requirement.

Solution: The duration of the video or audio file must be between xxx and yyy seconds.

400-InvalidFile.ImageSize

The size of image is beyond limit.

Cause: The image size exceeds the limit.

Solution: The image aspect ratio must not be greater than 2, and the longest side must not be greater than 4096.

400-InvalidFile.AspectRatio

Invalid file ratio. The file aspect ratio (height/width) must be between 3:1 and 1:3.

Cause: The file aspect ratio does not meet the requirement.

Solution: The aspect ratio of the video file must be between 3:1 and 1:3.

Invalid file ratio. The file aspect ratio (height/width) must be between 2.0 and 0.5.

Cause: The file aspect ratio does not meet the requirement.

Solution: The aspect ratio of the image file must be between 2.0 and 0.5.

400-InvalidFile.Openerror

Invalid file, cannot open file as video/audio/image.

Cause: The file cannot be opened.

Solution: Check whether the file is corrupted or whether the format is correct.

400-InvalidFile.Template.Content

Invalid template content.

Cause: You do not have permission for the action template, or the template content does not meet the requirements.

Solution: Check the template permissions and content.

400-InvalidFile.Format

Invalid file format，the request file format is one of the following types: MP4, AVI, MOV, MP3, WAV, AAC, JPEG, JPG, PNG, BMP, and WEBP.

Cause: The file format does not meet the requirement.

Solution: Use a file that meets the requirements. Supported video formats are mp4, avi, and mov. Supported audio formats are mp3, wav, and aac. Supported image formats are jpg, jpeg, png, bmp, and webp.

400-InvalidFile.MultiHuman

The input image has multi human bodies. Please upload other image with single person.

Cause: The input image has multiple people.

Solution: Upload a photo of a single person.

400-InvalidPerson

The input image has no human body or multi human bodies. Please upload other image with single person.

Cause: The input image has no person or multiple people.

Solution: Upload a photo of a single person.

400-FlowNotPublished

Flow has not published yet, please publish flow and try again.

Cause: The flow is not published.

Solution: Publish the flow and try again.

400-InvalidImage.ImageSize

The size of image is beyond limit.

Cause:s The image size exceeds the limit.

Solution: The image aspect ratio must not be greater than 2, and the longest side must not be greater than 4096.

400-InvalidImage.NoHumanFace

No human face detected.

Cause: No human face was detected. This error occurs only for asynchronous query APIs for generation tasks.

Solution: Upload an image that contains a clear human face.

400-InvalidImageResolution

The input image resolution is too large or small.

Cause: The input image resolution is too large or too small.

Solution: The image resolution must not be lower than 256 × 256 pixels and must not exceed 5760 × 3240 pixels.

400-InvalidImageFormat

The input image is in invalid format.

Cause: The image format does not meet the requirement.

Solution: Use an image in JPEG, PNG, JPG, BMP, or WEBP format.

400-InvalidURL

Invalid URL provided in your request.

Cause: The URL is invalid.

Solution: Use a valid URL.

Required URL is missing or invalid, please check the request URL.

Cause: The input URL is invalid or missing.

Solution: Provide the correct URL.

The request URL is invalid, make sure the url is correct and is an image.

Cause: The input URL is invalid.

Solution: Make sure that the URL is correct and points to an image file.

The input audio is longer than 20s.

Cause: The input audio file exceeds the maximum duration of 20 seconds.

Solution: Crop the audio file to be within 20 seconds.

File size is larger than 15MB.

Cause: The input audio file exceeds the maximum limit of 15 MB.

Solution: Compress the audio file to be within 15 MB.

File type is not supported. Allowed types are: .wav, .mp3.

Cause: The input audio format is not valid.

Solution: Currently, only wav and mp3 formats are supported.

The request URL is invalid, please check the request URL is available and the request image format is one of the following types: JPEG, JPG, PNG, BMP, and WEBP.

Cause: The image is not accessible or the downloaded file format is not supported.

Solution: Make sure that the URL is accessible and the image format is JPEG, JPG, PNG, BMP, or WEBP.

400-InvalidImage.FileFormat

Invalid image type. Please ensure the uploaded file is a valid image.

Cause: The image format is not supported.

Solution: Use a JPG, JPEG, PNG, BMP, or WEBP image.

400-InvalidURL.ConnectionRefused

Connection to xxx refused, please provide available URL.

Cause: The download was rejected.

Solution: Provide an available URL.

400-InvalidURL.Timeout

Download xxx timeout, please check network connection.

Cause: The download timed out.

Solution: Check your network connectivity.

400-BadRequestException

Invalid part type.

Cause: This occurs only in a Qwen-Long model conversation scenario when the user uploads a file type that is not currently supported by the Qwen-Long model.

Solution: Upload a file type that is supported by Qwen-Long.

400-BadRequest.EmptyInput

Required input parameter missing from request.

Cause: The input parameter was not added to the request.

Solution: Add the input parameter to the request.

400-BadRequest.EmptyParameters

Required parameter "parameters" missing from request.

Cause: The parameters parameter was not added to the request.

Solution: Add the parameters parameter to the request.

400-BadRequest.EmptyModel

Required parameter "model" missing from request.

Cause: The model parameter was not provided in the request.

Solution: Add the model parameter to the request.

400-BadRequest.IllegalInput

The input parameter requires json format.

Cause: The input parameter format does not conform to the JSON format required by the API.

Solution: Check the input parameter format to make sure that it is standard JSON.

400-BadRequest.InputDownloadFailed

Failed to download the input file: xxx.

Cause: Failed to download the input file, possibly due to a download timeout, download failure, or the file exceeding the size quota.

Solution: Troubleshoot based on the detailed error message xxx.

Failed to download the input file.

Cause: When using Qwen-TTS for voice cloning, the server failed to download the audio to be cloned.

Solution: Check whether the audio file can be downloaded, and whether its size exceeds the limit of 10 MB.

400-BadRequest.UnsupportedFileFormat

File format unsupported.

Cause: When you perform CosyVoice voice cloning, the uploaded audio format does not meet the model's requirements.

Solution: The audio format must be WAV (16-bit), MP3, or M4A. Note that you cannot judge the format by the file name extension alone. For example, a file with the .mp3 extension may be in another format, such as Opus. We recommend that you use a tool, such as ffprobe or mediainfo, or a command, such as the file command in Linux/macOS, to confirm the actual encoding format of the audio file to make sure that it meets the requirements.

Input file format is not supported.

Cause: The format of the input file is not supported.

Solution: Use a supported file format.

400-BadRequest.TooLarge

Payload Too Large.

Cause: The file size exceeds the limit.

Solution:

• When the "purpose" parameter is "file-extract", documents cannot exceed 150 MB and images cannot exceed 20 MB.

• When the "purpose" parameter is "batch", files cannot exceed 500 MB. Split the file and upload the files in batches.

400-BadRequest.ResourceNotExist

The Required resource not exist.

Cause:

• When you call the update, query, or delete API for CosyVoice voice cloning, the corresponding voice does not exist.

• When you use custom vocabulary (Paraformer), the custom vocabulary resource called by the update, query, or delete API does not exist.

400-Throttling.AllocationQuota

Your current quota is xxx

Cause: The number of CosyVoice voice cloning voices has reached the limit.

Solution: Delete some voices.

Free allocated quota exceeded.

Cause: When you use custom vocabulary (Paraformer), the number of custom vocabularies has exceeded the upper limit. The default is 10 per account.

Solution: Delete some custom vocabularies.

Maximum voice storage limit exceeded, please delete existing voices.

Cause: When using Qwen-TTS for voice cloning, your Alibaba Cloud account exceeds the voice number limit.

Solution: Delete some voices or request a quota increase.

400-InvalidGarment

Missing clothing image.Please input at least one top garment or bottom garment image.

Cause: A clothing image is missing.

Solution: Provide at least one image of a top garment (top_garment_url) or a bottom garment (bottom_garment_url).

400-InvalidSchema

Database schema is invalid for text2sql.

Cause: Database schema information was not entered.

Solution: Enter the database schema information.

400-InvalidSchemaFormat

Database schema format is invalid for text2sql.

Cause: The format of the input data table information is abnormal.

Solution: Check and correct the format of the data table information.

400-Audio.AudioShortError

valid audio too short!

Cause: The valid duration of the audio used for CosyVoice voice cloning is too short.

Solution:The audio duration should be controlled to be between 10 and 15 seconds. When you record, make sure to speak fluently and include at least one continuous speech segment of more than 5 seconds.

400-Audio.AudioSilentError

silent audio error.

Cause: The CosyVoice voice cloning audio file is silent or the non-silent length is too short.

Solution: The audio duration for voice cloning should be controlled to be between 10 and 15 seconds and include at least one continuous speech segment of more than 5 seconds.

400-InvalidInputLength

The image resolution is invalid, please make sure that the largest length of image is smaller than 4096, and the smallest length of image is larger than 150. and the size of image ranges from 5KB to 5MB.

Cause: The image dimensions or file size do not meet the requirements.

Solution: See Input image requirements.

400-FaqRuleBlocked

Input or output data is blocked by faq rule.

Cause: The FAQ rule intervention module was hit.

400-ClientDisconnect

Client disconnected before task finished!

Cause: The client actively disconnected before the task was finished. This error message is generated when you use speech synthesis or recognition services.

Solution: Check your code and do not disconnect from the server before the task is finished.

400-ServiceUnavailableError

Role must be user or assistant and Content length must be greater than 0.

Cause: The input content length is 0 or the role is incorrect.

Solution: Check that the input content length is greater than 0 and make sure that the parameter format, such as role, meets the requirements of the API reference.

400-IPInfringementSuspect

Input data is suspected of being involved in IP infringement.

Cause: The input data, such as the prompt or image, is suspected of intellectual property infringement.

Solution: Perform a content compliance check. Check the input to make sure that it does not contain content that poses an infringement risk.

400-UnsupportedOperation

The operation is unsupported on the referee object.

Cause: The associated object does not support the operation.

Solution: Check whether the operation object and operation type match.

The fine-tune job can not be deleted because it is succeeded,failed or canceled.

Cause: This fine-tuning task cannot be deleted because its status is already "succeeded", "failed", or "canceled".

Solution: Only tasks in specific states can be deleted. Do not delete tasks that are in a terminal state.

400-CustomRoleBlocked

Input or output data may contain inappropriate content with custom rule.

Cause: The request or response content did not pass the custom policy.

Solution: Check the content or adjust the custom policy.

400-Audio.PreprocessError

Audio preprocess error.

Cause: When using Qwen-TTS for voice cloning, the audio preprocessing failed. Possible reasons: The text parameter is significantly different from the audio, the effective human speech is too short, or the audio file contains no sound.

Solution: Adjust the text parameter. If the problem presists, refer to our recording guide and record the audio again.

No segments meet minimum duration requirement

Cause: When using Qwen-TTS for voice cloning, the effective human speech is too short in the audio to be cloned.

Solution: Refer to our recording guide and record the audio again.

400-BadRequest.VoiceNotFound

Voice '%s' not found.

Cause: When using Qwen-TTS for voice cloning, you call the delete voice interface, but the voice is already deleted or does not exist.

Solution: Check whether the voice parameter is correct.

400-Audio.DecoderError

Decoder audio file failed.

Cause: When using Qwen-TTS for voice cloning, the audio decoding failed. / CosyVoice voice cloning file decoding failed.

Solution: Check the audio file for corruption and ensure that it meets the required file format. For example, Qwen-TTS has specific requirements, while CosyVoice supports WAV (16-bit), MP3, or M4A.

400-Audio.AudioRateError

File sample rate unsupported.

Cause: When using Qwen-TTS or CosyVoice for voice cloning, the sample rate of the audio file is not supported.

Solution: Make sure the sample rate is at least 24000 Hz.

400-Audio.DurationLimitError

Audio duration exceeds maximum allowed limit.

Cause: When using Qwen-TTS for voice cloning, the audio to be cloned is too long.

Solution: Make sure the audio is no longer than 60 seconds.

401-InvalidApiKey/invalid_api_key

Invalid API-key provided. / Incorrect API key provided.

Cause: The API key is incorrect.

Solution: Common causes and their solutions are as follows:

• Incorrect environment variable

  • Incorrect: api_key=os.getenv("sk-xxx") . sk-xxx will be treated as the name of an environment variable, but not as a key.

  • Correct:

    • If you have exported the environment variable: Write as api_key=os.getenv("DASHSCOPE_API_KEY").

      Make sure you have set DASHSCOPE_API_KEY

    • If you havn't exported the environment variable: Write as api_key = "sk-xxx".

      Do not use this method in the production environment.

• Tool compatibility issues: A third-party tool is incompatible (For example, the latest version of the Dify plugin is unstable and may cause an error, you can try an older versio. When you use an older version of Cline, set the API Provider to OpenAI Compatible instead of Alibaba Qwen).

• Incorrect entry: API keys from Model Studio start with sk-. You may have entered an API key from another provider.

• Region mismatch: Your API key and base URL belong to different regions. For example, you may be using an API key from the China (Beijing) region with a base URL for the Singapore region. The base URL for the international region contains -intl. Verify that your API key is on the Singapore region page or the Beijing region page. The base URLs for each region are:

  Region	OpenAI compatible	DashScope
  International (Singapore)	https://dashscope-intl.aliyuncs.com/compatible-mode/v1	https://dashscope-intl.aliyuncs.com/api/v1
  China (Beijing)	https://dashscope.aliyuncs.com/compatible-mode/v1	https://dashscope.aliyuncs.com/api/v1

If none of the above apply, your API key may have been deleted. Obtain a new API key and try to call the API again.

401-NOT AUTHORIZED

Access denied: Either you are not authorized to access this workspace, or the workspace does not exist. Please:\nVerify the workspace configuration.\nCheck your API endpoint settings. Ensure you are targeting the correct environment.

Cause:

• The WorkspaceId value is invalid, or the current account is not a member of this workspace.

• The requested endpoint is incorrect.

Solution:

• Confirm that the WorkspaceId value is correct and that the account is a member of the workspace, then call the API again.

• Users on the China site must use the endpoint in the China (Beijing) region. Users on the international site must use the endpoint in the Singapore region. When you use API Debugging, confirm that the endpoint is correct (as shown in the following figures).

  

403-AccessDenied/access_denied

Current user api does not support asynchronous calls.

Cause: The API does not support asynchronous invocation.

Solution: Remove X-DashScope-Async from the request header, or set its value to disable.

current user api does not support synchronous calls.

Cause: The API does not support synchronous call.

Solution: Set X-DashScope-Async: enable in the request header.

Invalid according to Policy: Policy expired.

Cause: The file upload credential has expired when you obtain the temporary public URL.

Solution: Call the file upload credential API again to generate a new credential.

Access denied.

Reason: Access is denied because you do not have the required permissions for this model, or the model's free quota has been used up and it does not support paid usage (such as deepseek-r1-distill-llama-70b).

403-AccessDenied.Unpurchased

Access to model denied. Please make sure you are eligible for using the model.

Cause: You may not have activated the Alibaba Cloud Model Studio service yet.

Solution: Take the following steps to activate Model Studio service.

1. Register an account: If you do not have an Alibaba Cloud account, you must first register one.

2. Select a region: Model Studio provides the International Edition (Singapore) and the Mainland China Edition (Beijing). The two editions differ in terms of the console, endpoint, models, and their pricing, see Models.

   • International Edition: Endpoints and model services are located in international regions excluding Mainland China. The default endpoint is Singapore.

   • Mainland China Edition: Endpoints and model services are located in Mainland China. Currently, only the Beijing endpoint is supported.

3. Complete account verification: To activate the Mainland China Edition, this step is necessary. To activate the International Edition, you can skip this step.

   Use your Alibaba Cloud account to complete account verification:

   • Choose Individual Verification or Upgrade to Enterprise, and click Verify now.

   • In the Additional Information section of the page that appears, select Yes for Purchase cloud resources or enable acceleration inside the Chinese mainland.

   • For other items, see Account verification overview.

4. Activate Model Studio: Use your Alibaba Cloud account to go to the Model Studio console (Singapore or Beijing). After you read and agree to the Terms of Service, Model Studio is automatically activated. If the Terms of Service do not appear, this indicates that you have already activated Model Studio.

403-Model.AccessDenied

Model access denied.

Cause: You do not have permission to call the corresponding standard model.

Solution:

• Calling a standard model: When you use the API key of a sub-workspace to call a standard model, such as qwen-plus, the sub-workspace must have permission to call that model. For more information, see Authorize a sub-workspace to call models.

403-App.AccessDenied

App access denied.

Cause: You do not have permission to access the application or model.

Solution:

• Carefully confirm that access authorization has been granted to the workspace and RAM user.

• Carefully check whether the application is published.

• Carefully verify that the APP ID and API KEY passed in are correct.

• If the Claude Code reports an error, use the API key of the default workspace.

• If all the above suggestions are correct, we recommend that you refresh the data, republish, and then call again, or try to create the agent again.

403-Workspace.AccessDenied

Workspace access denied.

Cause: You do not have permission to access the application or model of the workspace.

Solution:

• If you are calling a model in a sub-workspace, see Call a model in a sub-workspace.

• You can also use the API key of the Alibaba Cloud account, which has permissions for all workspaces.

403-AllocationQuota.FreeTierOnly

The free tier of the model has been exhausted. If you want to continue access the model on a paid basis, please disable the "use free tier only" mode in the management console.

Cause: The Stop when free quota is used up feature is enabled, and a request was sent after the free quota was exhausted.

There is an hourly delay in the display of the free quota in the console. Even if the quota has been used up, the interface may still show a remaining quota.

Solution: To make a paid call, wait for the console to show that the free quota is used up, then turn off the Stop when free quota is used up button.

404-ModelNotFound/model_not_found

The provided model xxx is not supported by the Batch API.

Cause: The current model does not support Batch calls, or the model name may be misspelled.

Solution: For more information, see OpenAI compatible - Batch to confirm the models that support Batch calls and their correct names.

Model can not be found. / The model xxx does not exist. / The model xxx does not exist or you do not have access to it.

Cause: The model being accessed does not exist, or you have not yet activated the Model Studio service.

Solution:

• Check the model name in Models to verify that the model name you entered (the value of the model parameter) is correct.

• Go to the Models page in the console to activate the model service.

404-model_not_supported

Unsupported model xxx for OpenAI compatibility mode.

Cause: The current model does not support access in OpenAI compatible mode.

Solution: Call the model using the DashScope method.

404-WorkSpaceNotFound

WorkSpace can not be found.

Cause: The workspace does not exist.

404-NotFound

Not found!

Cause:

• The resource to be queried or operated on does not exist.

• When you use custom vocabulary, the passed custom vocabulary ID is invalid or the corresponding custom vocabulary does not exist.

Solution:

• Check whether the ID of the resource to be queried or operated on is incorrect.

• Check whether the custom vocabulary ID is correct and call it in the correct way according to the API reference.

429-Throttling

Requests throttling triggered.

Cause: The API call triggered rate limiting.

Solution: Reduce the call frequency or try again later.

Too many requests in route. Please try again later.

Cause: Too many requests triggered rate limiting.

Solution: Please try again later.

429-Throttling.RateQuota/LimitRequests/limit_requests

Requests rate limit exceeded, please try again later. /You exceeded your current requests list.

Cause: The call frequency (records per second (RPS)/RPM Package Manager (RPM)) triggered rate limiting.

Solution: For more information, see Rate limits and control the call frequency.

429-Throttling.BurstRate/limit_burst_rate

Request rate increased too quickly. To ensure system stability, please adjust your client logic to scale requests more smoothly over time.

Cause: When the rate limit was not met, the call frequency increased sharply, which triggered the system stability protection mechanism.

Solution: We recommend that you optimize the client call logic and adopt a smooth request strategy, such as uniform scheduling, exponential backoff, or request queue buffering, to distribute requests evenly within the time window and avoid instantaneous peaks.

429-Throttling.AllocationQuota/insufficient_quota

Allocated quota exceeded, please increase your quota limit./ You exceeded your current quota, please check your plan and billing details.

Cause: The number of tokens consumed per second or per minute (TPS/TPM) triggered Rate limiting.

Solution: For more information, see How are tokens calculated? and adjust the call strategy according to the Rate limiting rules.

Too many requests. Batch requests are being throttled due to system capacity limits. Please try again later.

Cause: Too many batch requests triggered rate limiting.

Solution: Your request cannot be processed at this time. Please try again later.

Free allocated quota exceeded.

Cause: The free quota has expired or been used up, and the model does not currently support the pay-as-you-go billing method.

Solution: Use another model as a replacement. For example, if the Qwen-Audio model quota is used up, you can use the omni-modal model.

429-CommodityNotPurchased

Commodity has not purchased yet.

Cause: The workspace has not been subscribed to.

Solution: Subscribe to the workspace service first.

429-PrepaidBillOverdue

The prepaid bill is overdue.

Cause: The subscription bill for the workspace has expired.

429-PostpaidBillOverdue

The postpaid bill is overdue.

Cause: The model inference product has expired.

430-Audio.DecoderError

Decoder audio file failed.

Cause: The CosyVoice voice cloning audio file failed to decode.

Solution: We recommend that you use a tool, such as ffprobe or mediainfo, or a command, such as the file command in Linux/macOS, to confirm the actual encoding format of the audio file to make sure that it meets the requirements.

430-Audio.FileSizeExceed

File too large

Cause: The CosyVoice voice cloning audio file size exceeds the limit.

Solution: The audio file for voice cloning must be within 10 MB.

430-Audio.AudioRateError

File sample rate unsupported

Cause: The CosyVoice voice cloning audio file sample rate is not supported.

Solution: Set the sample rate to 16 kHz or higher.

430-Audio.AudioSilentError

Silent file unsupported.

Cause: The CosyVoice voice cloning audio file is silent or the non-silent length is too short.

Solution: The audio duration should be controlled to be between 10 and 15 seconds and include at least one continuous speech segment of more than 5 seconds.

500-InternalError/internal_error

An internal error has occured, please try again later or contact service support.

Cause: An internal error occurred.

Solution:

• If you are using the Qwen-Omni model, you need to use the streaming output method.

• If you are using CosyVoice voice cloning, possible causes are:

  • The audio file is not standard. For example, there is a problem with the sound itself, there is noise, or the sound fluctuates. Refer to Recording guide and retry.

  • The recording file URL is inaccessible. Follow the instructions in the CosyVoice voice cloning API and retry.

  • The recording file is too long. Try to choose a recording of 10 to 15 seconds. When you record, make sure to speak fluently and include at least one continuous speech segment of more than 5 seconds.

Internal server error!

Cause: An internal algorithm error occurred.

Solution: Please try again later.

audio preprocess server error

When you use CosyVoice voice cloning:

• Cause:The audio file is not standard. For example, there is a problem with the sound itself, there is noise, or the sound fluctuates.

  Solution: For more information, see the Recording guide. Record again, and retry.

• Cause: The recording file URL is inaccessible.

  Solution: Follow the instructions in the CosyVoice Voice cloning API and retry.

• Cause: The recording file is too long.

  Solution: Try to choose a recording of 10 to 15 seconds. When you record, make sure to speak fluently and include at least one continuous speech segment of more than 5 seconds.

500-InternalError.FileUpload

oss upload error.

Cause: File upload failed.

Solution: Check your OSS configuration and network.

500-InternalError.Upload

Failed to upload result.

Cause: Failed to upload the generation result.

Solution: Check the storage configuration or try again later.

500-InternalError.Algo

inference internal error.

Cause: A service exception occurred.

Solution: Try again first to rule out a sporadic issue.

Expecting ',' delimiter: line x column xxx (char xxx)

Cause: The JSON data generated by the model is invalid, and a tool call cannot be initiated normally.

Solution: Switch to the latest model or optimize the prompt and try again.

Missing Content-Length of multimodal url.

Cause: The Content-Length field is missing from the response header of the URL request.

Solution: If the problem cannot be solved, try to use another image link.

An error occurred in model serving, error message is: [Request rejected by inference engine!]

Cause: An error occurred on the underlying server of the model service.

Solution: Please try again later.

An internal error has occured during algorithm execution.

Cause: An error occurred during algorithm runtime.

Solution: Please try again later.

Inference error: Inference error.

Cause: An inference error occurred.

Solution: Check whether the input image file is corrupted or check the quality of the person image. It must contain a complete and clear face.

Role must be in [user, assistant]

Cause: When you use the Qwen-MT model, the messages array contains a message with a role other than user.

Solution: Make sure that the messages array contains only one element, and that element must be a User Message.

Embedding_pipeline_Error: xxx

Cause: An error occurred in image or video pre-processing.

Solution: Confirm that the uploaded image or video and the request code meet the requirements, then try again.

Receive batching backend response failed!

Cause: An internal service error occurred.

Solution: Please try again later.

An internal error has occured during execution, please try again later or contact service support. / algorithm process error. / inference error. / An internal error occurs during computation, please try this model later.

Cause: An internal algorithm error occurred.

Solution: Please try again later.

list index out of range

Cause: The last element in the messages array must be a User Message.

Solution: Adjust the order of the messages array to make sure that the last element is {"role": "user", ...}.

500-InternalError.Timeout

An internal timeout error has occured during execution, please try again later or contact service support.

Cause: After an asynchronous task was submitted, no result was returned within 3 hours, which caused a timeout.

Solution: Check the task execution status or contact technical support.

500-SystemError

An system error has occured, please try again later.

Cause: A system error occurred.

Solution: Please try again later.

500-ModelServiceFailed

Failed to request model service.

Cause: The model service call failed.

Solution: Please try again later.

500-RequestTimeOut

Request timed out, please try again later. / Response timeout! / I/O error on POST request for "https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions": timeout

Cause:

• The request timed out when you call the large language model. The timeout error time is 300 seconds.

• When you use speech recognition (Paraformer), no audio was sent to the server for a long time, or silent audio was sent for a long time.

Solution:

• Send the request using streaming output. For more information, see Streaming.

• Set the request parameter heartbeat to true or end the recognition task promptly.

500-InvokePluginFailed

Failed to invoke plugin.

Cause: The plug-in call failed.

Solution: Check the plug-in configuration and availability.

500-AppProcessFailed

Failed to proceed application request.

Cause: The application flow processing failed.

Solution: Check the application configuration and flow nodes.

500-RewriteFailed

Failed to rewrite content for prompt.

Cause: The call to the prompt rewriting model failed.

Solution: Please try again later.

500-RetrievalFailed

Failed to retrieve data from documents.

Cause: Document retrieval failed.

Solution: Check the document index and retrieval configuration.

503-ModelServingError

Too many requests. Your requests are being throttled due to system capacity limits. Please try again later.

Cause: Network resources are currently saturated, and your request cannot be processed at this time.

Solution: Please try again later.

503-ModelUnavailable

Model is unavailable, please try again later.

Cause: The model is temporarily unavailable.

Solution: Please try again later.

SDK errors

error.AuthenticationError: No api key provided. You can set by dashscope.api_key = your_api_key in code, or you can set it via environment variable DASHSCOPE_API_KEY= your_api_key.

Cause: An API key was not provided when you use the DashScope SDK.

Solution: For information about how to configure an API key, see Set an API key as an environment variable.

openai.OpenAIError: The api_key client option must be set either by passing api_key to the client or by setting the OPENAI_API_KEY environment variable

Cause: An API key was not passed.

Solution:

• Pass the API key using an environment variable (Recommended)

  Set the DASHSCOPE_API_KEY environment variable (see Set API key as environment variable). When you initialize the client, read the key using the os.getenv method:

  client = OpenAI(api_key=os.getenv("DASHSCOPE_API_KEY"),...)

• Pass the API key in plaintext (For testing only)

  Pass the API key directly to the api_key parameter:

  client = OpenAI(api_key="sk-...", ...)

  Note: This method poses a security risk. Do not use it in a production environment.

Bad Request for url: xxx

Cause: When you use the Python requests library, adding the response.raise_for_status() statement prevents the specific error content from the server from being returned when an error occurs.

Solution: Use print(response.json()) to view the information returned by the server.

Cannot resolve symbol 'ttsv2'

Cause: If you are using speech synthesis (CosyVoice), this problem is caused by an outdated version of the DashScope SDK.

Solution: Install the latest version of the DashScope SDK.

NetworkError

NoApiKeyException: Can not find api-key.

Cause: The environment variable has not taken effect.

Solution: You can restart the client or IDE and try again. For more information, see FAQ.

ConnectException: Failed to connect to dashscope.aliyuncs.com

Cause: An exception exists in the on-premises network environment.

Solution: Check your on-premises network. For example, you may be unable to access HTTPS due to a certificate issue, or the firewall may be set incorrectly. We recommend that you switch to a different network environment or server for testing.

InputRequiredException: Parameter invalid: text is null

Cause: When you use speech synthesis (CosyVoice), the text to be synthesized was not sent.

Solution: Assign a value to the text parameter when you call the speech synthesis API.

MultiModalConversation.call() missing 1 required positional argument: 'messages'

Cause: The DashScope SDK version is outdated.

Solution: Install the latest version of the DashScope SDK.

mismatched_model

The model 'xxx' for this request does not match the rest of the batch. Each batch must contain requests for a single model.

Cause: In a single batch task, all requests must use the same model.

Solution: Check your input file according to the input file format.

duplicate_custom_id

The custom_id 'xxx' for this request is a duplicate of another request. The custom_id parameter must be unique for each request in a batch.

Cause: In a single batch task, the ID of each request must be unique.

Solution: Check your input file according to the input file format and make sure that all request IDs are unique.

Upload file capacity exceed limit. / Upload file number exceed limit.

Cause: The file upload failed. The Model Studio storage space under the current Alibaba Cloud account is full or nearly full.

Solution: You can use the OpenAI compatible - File API to delete unnecessary files to free up space. The current storage space supports a maximum of 10,000 files and a total of no more than 100 GB.

WebSocket errors

The decoded text message was too big for the output buffer and the endpoint does not support partial messages

Cause: When you use streaming speech recognition with speech recognition (Paraformer), the amount of recognition result data returned by the service is too large.

Solution: Send the audio to be recognized in segments. We recommend that the duration of the audio sent each time be about 100 milliseconds, and the data size be kept between 1 KB and 16 KB.

TimeoutError: websocket connection could not established within 5s. Please check your network connection, firewall settings, or server status.

Cause: If you are using speech synthesis (CosyVoice), a WebSocket connection could not be established within 5 seconds.

Solution: Check your on-premises network and firewall settings, or switch to a different network environment or server for testing.

unsupported audio format:xxx

Cause: When you perform CosyVoice voice cloning, the uploaded audio format does not meet the model's requirements.

Solution: The audio format must be WAV (16-bit), MP3, or M4A. Note that you cannot judge the format by the file name extension alone. We recommend that you use a tool, such as ffprobe or mediainfo, or a command, such as the file command in Linux/macOS, to confirm the actual encoding format of the audio file.

internal unknown error

Cause: The CosyVoice voice cloning audio file format may not meet the requirements.

Solution: The audio format must be WAV (16-bit), MP3, or M4A. We recommend that you use a tool, such as ffprobe or mediainfo, to confirm the actual encoding format of the audio file.

Invalid backend response received (missing status name)

Cause: When you use the RESTful API for audio file recognition with speech recognition (Paraformer), a request parameter is misspelled.

Solution: Refer to the API reference to check your code.

NO_INPUT_AUDIO_ERROR

Cause: No valid speech was detected.

Solution: If you are using real-time speech recognition with speech recognition (Paraformer), troubleshoot as follows:

1. Check whether there is audio input.

2. Check whether the audio format is correct. pcm, wav, mp3, opus, speex, aac, and amr are supported.

SUCCESS_WITH_NO_VALID_FRAGMENT

Cause: If you are using audio file recognition with speech recognition (Paraformer), the recognition result query API was called successfully, but the VAD module did not detect valid speech.

Solution: Check whether the audio file contains valid speech. If it is all invalid speech, such as pure silence, it is normal to have no recognition result.

ASR_RESPONSE_HAVE_NO_WORDS

Cause: If you are using audio file recognition with speech recognition (Paraformer), the recognition result query API was called successfully, but the final recognition result is empty.

Solution: Check whether the audio file contains valid speech, or whether the valid speech is all filler words and the disfluency detection parameter disfluency_removal_enabled is enabled, which causes the filler words to be filtered out.

FILE_DOWNLOAD_FAILED

Cause: If you are using audio file recognition with speech recognition (Paraformer), the file to be recognized failed to download.

Solution: Check whether the audio file path is correct and whether it can be accessed and downloaded from the external network.

FILE_CHECK_FAILED

Cause: If you are using audio file recognition with speech recognition (Paraformer), the file format is incorrect.

Solution: Check whether the audio file is in single-track/dual-track WAV format or MP3 format.

FILE_TOO_LARGE

Cause: If you are using audio file recognition with speech recognition (Paraformer), the file to be recognized is too large.

Solution: Check whether the audio file size exceeds 2 GB. If it does, you need to segment the audio file.

FILE_NORMALIZE_FAILED

Cause: If you are using audio file recognition with speech recognition (Paraformer), the file to be recognized failed to normalize.

Solution: Check whether the audio file is corrupted and whether it can be played normally.

FILE_PARSE_FAILED

Cause: If you are using audio file recognition with speech recognition (Paraformer), the file failed to parse.

Solution: Check whether the audio file is corrupted and whether it can be played normally.

MKV_PARSE_FAILED

Cause: If you are using audio file recognition with speech recognition (Paraformer), the MKV failed to parse.

Solution: Check whether the audio file is corrupted and whether it can be played normally.

FILE_TRANS_TASK_EXPIRED

Cause: If you are using audio file recognition with speech recognition (Paraformer), the audio file recognition task has expired.

Solution: The TaskId does not exist or has expired. Resubmit the task.

REQUEST_INVALID_FILE_URL_VALUE

Cause: If you are using audio file recognition with speech recognition (Paraformer), the file_link parameter of the request is invalid.

Solution: Confirm that the format of the file_url parameter is correct.

CONTENT_LENGTH_CHECK_FAILED

Cause: If you are using audio file recognition with speech recognition (Paraformer), the content-length check failed.

Solution: When you download the audio file to be recognized, check whether the content-length in the HTTP response is consistent with the actual file size.

FILE_404_NOT_FOUND

Cause: If you are using audio file recognition with speech recognition (Paraformer), the file to be downloaded does not exist.

Solution: Check whether the file URL is correct.

FILE_403_FORBIDDEN

Cause: If you are using audio file recognition with speech recognition (Paraformer), you do not have permission to download the audio to be recognized.

Solution: Check the file access permissions.

FILE_SERVER_ERROR

Cause: If you are using audio file recognition with speech recognition (Paraformer), the service where the requested file is located is unavailable.

Solution: Try again later or check the file server status.

AUDIO_DURATION_TOO_LONG

Cause: If you are using audio file recognition with speech recognition (Paraformer), the duration of the requested file exceeds 12 hours.

Solution: We recommend that you segment the audio and submit the recognition tasks multiple times. You can use tools such as FFmpeg to segment the audio.

DECODE_ERROR

Cause: If you are using audio file recognition with speech recognition (Paraformer), it failed to detect the audio file information.

Solution: Confirm that the file in the file download link is in a supported audio format.

CLIENT_ERROR-[qwen-tts:]Engine return error code: 411

Cause: When you perform Qwen-TTS real-time speech synthesis, the selected model is qwen-tts-vc-realtime-2025-08-20, but the voice is the default voice. This model supports only cloned voices.

Solution: Use a voice generated by voice cloning, not the default voice.

NO_VALID_AUDIO_ERROR

Cause: When you use speech recognition (Paraformer), the audio to be recognized is invalid. Solution: Check whether the audio format, sample rate, and other parameters meet the requirements.