Alibaba Cloud Model Studio:Error codes

400-InvalidParameter

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

Cause: You called a thinking-mode model using a non-streaming output method.

Solution: Set the enable_thinking parameter to false, or use the streaming output method to call the thinking-mode model.

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

Cause: The thinking_budget parameter is outside the allowed range.

Solution: Refer to the maximum chain-of-thought length listed for the model in the Model List and set this parameter to a value greater than 0 and no larger than that limit.

This model only support stream mode, please enable the stream parameter to access the model. / current user api does not support http call.

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

Solution: Use the streaming output method to call the model.

This model does not support enable_search.

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

Solution: Call a model that supports web search.

Current language settings are not supported!

Cause: When using the Qwen-MT model, the source_lang or target_lang format is incorrect or not included in the supported languages.

Solution: Provide the correct English name or language code.

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

Cause: When the thinking mode is enabled, 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 before calling. The API will then return 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 before calling. The API will then return incremental content.

Range of input length should be [1, xxx]

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

Solution:

• If calling via code, ensure the total token count in the messages array stays within the model’s maximum input token limit.

• When using a chat client (such as Chatbox) or the Alibaba Cloud Model Studio console for continuous conversations, each request includes conversation history, which can easily exceed the model’s limit. If this happens, start a new conversation.

Range of max_tokens should be [1, xxx]

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

Solution: Refer to the "Maximum Output Tokens" value in the Model List documentation for the upper limit of max_tokens.

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

Cause: The temperature parameter is outside the [0.0, 2.0) range.

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 top_p parameter is outside the (0.0, 1.0] range.

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 presence_penalty parameter is outside the [-2.0,2.0] range.

Solution: Set the presence_penalty parameter within the [-2.0,2.0] range.

Range of n should be [1, 4]

Cause: The n parameter is not within the [1, 4] range.

Solution: Set the n parameter within the [1, 4] range.

Range of seed should be [0, 9223372036854775807]

Cause: When using the DashScope protocol, the seed parameter is not within the [0, 9223372036854775807] range.

Solution: Set the seed parameter within the [0, 9223372036854775807] range.

Request method 'GET' is not supported.

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

Solution: Consult the API reference and use a supported request method (such as POST) to resend the request.

messages with role "tool" must be a response to a preceeding message with "tool_calls"

Cause: During tool calling, you did not add an Assistant Message to the messages array.

Solution: Add the model’s first Assistant Message response to the messages array before adding the Tool Message.

Required body invalid, please check the request body format.

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

Solution: Check the request body to ensure it is valid JSON. Common issues include trailing commas (,) or unclosed brackets. You can use the large model to help fix the request body format.

input content must be a string.

Cause: Plain-text models do not support setting the content field in messages to a non-string type.

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

The content field is a required field.

Cause: When making a request, you did not specify the content parameter, for example: {"role": "user"}.

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

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

Cause: When calling a large language model, you specified neither the messages parameter nor the prompt parameter (which is deprecated). If you specified messages but still received an error, the format might be incorrect—for example, when using DashScope-HTTP, messages must be placed inside an input object, not at the same level as the model parameter.

Solution: Specify the messages parameter. If you already did but still get an error, see the text generation API reference to verify its placement.

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

Cause: When using structured output, the prompt does not include the keyword json.

Solution: Add the keyword json (case-insensitive) to your prompt, for example: “Please output in JSON format.”

Json mode response is not supported when enable_thinking is true

Cause: You enabled the model’s thinking mode while using structured output.

Solution: When using structured output, set enable_thinking to false to disable thinking mode. Alternatively, see the FAQ How to use structured output with thinking-mode models?

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 does not comply with requirements.

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

The value of the enable_thinking parameter is restricted to True.

Cause: For certain models (such as qwen3-235b-a22b-thinking-2507), you cannot set the enable_thinking parameter to false.

Solution:

• If calling through a third-party tool (such as Cherry Studio), enable the thinking toggle in the input box.

• If calling via code, set enable_thinking to true.

'audio' output only support with stream=true

Cause: When using the Qwen-Omni model, you did not use streaming output, but the model supports only streaming.

Solution: Set the stream parameter to true to enable streaming output.

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

Cause: The tool_choice parameter is incorrect during Function Calling.

Solution: Set it to "auto" (let the LLM choose tools autonomously) or "none" (force no tool usage).

Model not exist.

Cause: The model parameter is invalid or incorrectly formatted.

Solution:

• Check model name format: Verify that the model parameter uses correct capitalization and contains no extra spaces.

• Use the correct model name: Check the model you entered against the model names in the Model List to ensure it is correct. Do not mix model names from the open-source community with Model Studio model IDs, such as using 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 calling 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, resulting in insufficient sampling points.

Solution: Increase the audio duration.

File parsing in progress, please try again later.

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

Solution: Wait for file parsing to complete before retrying.

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 required format: str, list[str], list[int], or list[list[int]].

Solution: See the text generation API reference and set the stop parameter correctly.

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

Cause: When calling an Embedding model, the number of texts exceeds the model’s limit.

Solution: Refer to the batch size information in the Embedding documentation and control the number of input texts.

[] is too short

Cause: The input messages array is empty.

Solution: Add a message before sending the request.

The tool call is not supported.

Cause: The model you are using does not support the tools parameter.

Solution: Switch to a Qwen or DeepSeek model that supports Function Calling.

Repetitive tool calls detected in the conversation history. The same tool call with identical name and arguments has been repeated across multiple consecutive rounds. Please modify your request or adjust the tool call arguments to avoid infinite loops.

Cause: Repetitive tool calls were detected in the conversation history: a tool call with the identical name and arguments was repeated across multiple consecutive rounds, indicating the model may be stuck in a loop. The HTTP status code is 400 and the error code isInternalError.Algo.InvalidParameter.

Solution:

• After each round of tool calls, append the tool execution result as a Tool Message to themessagesarray before sending the next request, so the model can proceed based on the result instead of repeating the same call.

• Check whether the tool returns results correctly. If the tool keeps returning the same or invalid results, add a maximum call limit or termination logic on the application side to avoid infinite loops.

• If needed, refine the prompt to clearly define the task completion conditions and prevent the model from repeatedly issuing the same tool call.

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

Cause: The API call parameters are invalid.

Solution: Check the request parameters to ensure all required parameters are provided and correctly formatted.

When the error message specifies the parameter data_sources, it usually means that when calling the CreateIndex API, the required parameter SourceType was omitted, causing the subsequent SubmitIndexJob API to fail. When creating a knowledge base from documents, set this parameter to DATA_CENTER_FILE; when creating from categories, set it to DATA_CENTER_CATEGORY. See the CreateIndex documentation for details.

input must contain file_urls

Cause: When using speech recognition (Paraformer) for recorded file recognition, you did not assign a value to the file_urls request parameter.

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

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

Cause: When using visual understanding, omni-modal, or audio understanding models, the provided URL or local path is invalid or does not meet requirements.

Solution:

• Passing a URL: It must start with http://, https://, or data:. If starting with data:, include "base64" before the Base64-encoded data.

• Passing a local path: It must start with file://.

• Passing a temporary URL:

  • For HTTP calls, ensure the request header includes X-DashScope-OssResourceResolve: enable.

  • For SDK calls: Only DashScope SDK is supported—do not use the OpenAI SDK.

Input should be a valid dictionary or instance of GPT3Message

Cause: The messages field format is invalid—for example, mismatched brackets or missing key-value pairs.

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 using an Embedding model, the input is neither a string nor a list of strings.

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

The video modality input does not meet the requirements because: the range of sequence images shoule be (4, 512)./(4,80).

Cause: When using the Qwen VL model to input video as a list of images, the number of images does not meet requirements.

Solution: For Qwen3-VL and Qwen2.5-VL series models, provide 4–512 images; for other models, provide 4–80 images. See Image and Video Understanding for details.

Exceeded limit on max bytes per data-uri item : 10485760'. / Multimodal file size is too large

Cause: The local image or video file passed to a multimodal model (Qwen-VL, QVQ, Qwen-Omni) exceeds the size limit.

Solution:

• Local files: After Base64 encoding, a single file must not exceed 10 MB.

• File URLs: Image files must not exceed 10 MB; for video files:

  • Qwen3-VL, qwen-vl-max: up to 2 GB;

  • qwen-vl-plus series: up to 1 GB;

  • Other models: up to 150 MB.

See How to compress images or videos to meet size requirements?

Input should be 'Cherry', 'Serena', 'Ethan' or 'Chelsie': parameters.audio.voice

Cause: When using Qwen-Omni or Qwen-TTS, the voice parameter is incorrect.

Solution: Set it to one of: 'Cherry', 'Serena', 'Ethan', or 'Chelsie'.

The image length and width do not meet the model restrictions.

Cause: The dimensions (width and height) of the image passed to the Qwen VL model do not meet model requirements.

Solution: Image dimensions must satisfy: both width and height are at least 10 pixels, and the aspect ratio must not exceed 200:1 or 1:200.

Failed to decode the image during the data inspection.

Cause: Image decoding failed.

Solution: Confirm the image is not corrupted and that its format is supported.

The file format is illegal and cannot be opened. / The audio format is illegal and cannot be opened. / The media format is not supported or incorrect for the data inspection.

Cause: The file format is unsupported or the file cannot be opened.

Solution: Confirm the file is not corrupted, that the file name extension matches the actual format, and that the format is supported.

The input messages do not contain elements with the role of user.

Cause:

• When calling the model, you did not pass a User Message;

• Or, when calling an Alibaba Cloud Model Studio workflow application via API, parameters passed to the start node must be sent via the biz_params parameter (not user_prompt_params).

Solution: Ensure you pass a User Message to the model or correctly pass custom parameters.

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: The server cannot download the media file from the public URL, possibly due to:

• Connectivity issue: Using an internal network address from Alibaba Cloud Object Storage Service.

• Network latency: Cross-region access causing timeouts.

• Service instability: Slow or unreachable source storage service.

Solution:

• Change storage service

  Use a storage service in the same region as the model service. We recommend using Alibaba Cloud Object Storage Service to generate public URLs (do not use internal addresses).

• Adjust transfer method

  If passing a public URL fails, refer to Pass local files (Base64 encoded or file paths) and switch to the recommended method:

  File type	File specs	DashScope SDK (Python, Java)	OpenAI compatible / DashScope HTTP
  Image	Larger than 7 MB but smaller than 10 MB	Pass local path	Public URL only; we recommend using Alibaba Cloud Object Storage Service
  	Smaller than 7 MB	Pass local path	Base64 encoded
  Video	Larger than 100 MB	Public URL only; we recommend using Alibaba Cloud Object Storage Service	Public URL only; we recommend using Alibaba Cloud Object Storage Service
  	Larger than 7 MB but smaller than 100 MB	Pass local path	Public URL only; we recommend using Alibaba Cloud Object Storage Service
  	Smaller than 7 MB	Pass local path	Base64 encoded
  Audio	Larger than 10 MB	Public URL only; we recommend using Alibaba Cloud Object Storage Service	Public URL only; we recommend using Alibaba Cloud Object Storage Service
  	Larger than 7 MB but smaller than 10 MB	Pass local path	Public URL only; we recommend using Alibaba Cloud Object Storage Service
  	Smaller than 7 MB	Pass local path	Base64 encoded

  Base64 encoding increases data size; the original file should be smaller than 7 MB.

  Using Base64 or local paths avoids server-side download timeouts and improves stability.

Failed to find the requested media resource during the data inspection process.

Cause: The provided resource URL is invalid or inaccessible.

Solution:

• Confirm the URL is correctly formatted and accessible.

• Confirm the resource file has not been deleted or moved.

• Confirm the URL has not expired (for OSS signed URLs, check validity period).

url error, please check url！

• Cause 1: Model name does not match the API endpoint: For example, using a multimodal endpoint for a plain-text model, or a plain-text endpoint for a multimodal model.

  Solution:

  1. When using multimodal models like qwen3.7-plus or qwen3-vl-plus via DashScope: use MultiModalConversation.call() or the multimodal-generation endpoint. See Image and Video Understanding.

     If using the spring-ai-alibaba framework, confirm you have set the multimodal parameter withMultiModel.

  2. When using plain-text models like qwen3-max, qwen-plus, or deepseek-v4-pro via DashScope, use Generation.call() or the text-generation endpoint. See Overview.

  3. When using the CosyVoice voice cloning API via DashScope: this API includes model and target_model parameters. Set model to voice-enrollment and target_model to the specific CosyVoice model. See CosyVoice Voice Cloning/Design API.

• Cause 2: DashScope SDK version is outdated: Older SDK versions cannot recognize the correct server address when calling image/video generation models.

  Solution: Upgrade the SDK version

Don't have authorization to access the media resource during the data inspection process.

Cause: The signed OSS file URL passed during model invocation has expired.

Solution: Ensure you access the file within the URL’s validity period.

The item of content should be a message of a certain modal.

Cause: When using the DashScope SDK to call a multimodal model, each element in the content array must have a key that is one of: image, video, audio, or text.

Solution: Use the correct content parameter.

Invalid video file.

Cause: The provided video file is invalid.

Solution: Check if the video file is corrupted or if the format is correct.

The video modality input does not meet the requirements because: The video file is too long.

Cause: The video duration exceeds the limit for Qwen VL or Qwen-Omni models.

Solution:

• Qwen2.5-VL supports videos between 2 seconds and 10 minutes.

• Other Qwen VL or Qwen-Omni models support videos between 2 seconds and 40 seconds.

Field required: xxx

Cause: A required input parameter is missing.

Solution: Add the missing parameter as indicated by the error message xxx.

The request is missing required parameters or in a wrong format, please check the parameters that you send.

Cause: Required parameters are missing or incorrectly formatted.

Solution: Check that all request parameters are complete and correctly formatted.

Invalid ext_bbox.

Cause: The input ext_bbox is invalid.

Solution: See Emoji Video Generation for details.

Driven not exist: driven_id.

Cause: The input driven_id does not exist.

Solution: See Emoji Video Generation for details.

Missing training files.

Cause: Parameter error—missing parameters or incorrect formatting.

The style is invalid.

Cause: The style value is outside the allowed enumeration.

Solution: Check that the style parameter value is correct.

The style_level is invalid.

Cause: The style_level value is outside the allowed enumeration.

Solution: See EMO Video Generation for details.

parameters.video_ratio must be 9:16 or 3:4.

Cause: The video_ratio parameter can only be 9:16 or 3:4.

Solution: Set the video_ratio parameter to "9:16" or "3:4".

the xxx parm is invalid!

Cause: Input parameter exceeds allowed range.

Solution: See Video Style Redrawing for details.

input json error.

Cause: Input JSON is invalid.

Solution: Check that the request JSON format is correct.

read image error.

Cause: Failed to read the image.

Solution: Check if the image file is corrupted or if the format is correct.

the parameters must conform to the specification: xxx.

Cause: Input parameter value exceeds allowed range.

Solution: Check and correct the parameter value as indicated by the error message xxx.

The size of person image and coarse_image are not the same.

Cause: The resolution of coarse_image does not match person_image.

Solution: Ensure the resolutions of coarse_image and person_image are identical.

The request is missing required parameters or the parameters are out of the specified range, please check the parameters that you send.

Cause: Required API parameters are missing or out of range.

Solution: Check and correct the request parameters.

image format error

Cause: Image format is incorrect.

Solution: Must be an image URL or Base64 string.

No messages found in input

Cause: The request parameters must include a messages field.

Solution: See Qwen - Image Editing for details.

Invalid image format or corrupted file

Cause: Input image format is incorrect or the file is corrupted.

Solution: Check that the file can be opened and downloaded normally, and ensure it is complete and in a supported format.

download image failed

Cause: The image cannot be downloaded.

Solution: Check that the file can be downloaded normally.

messages length only support 1

Cause: The messages array length supports only 1.

Solution: Only one conversation message can be passed. See Qwen - Image Editing for details.

content length only support 2

Cause: The content array length supports only 2.

Solution: Only one text and one image can be passed. See Qwen - Image Editing for details.

lack of image or text

Cause: The request parameters are missing the image or text field.

Solution: See Qwen - Image Editing for details.

num_images_per_prompt must be 1.

Cause: The request parameter is invalid; the parameter n (number of images to generate) can only be set to 1.

Solution: Set the value of parameter n to 1.

Input files format not supported.

Cause: Audio or image format does not meet requirements.

Solution: Supported audio formats: mp3, wav, aac; supported image formats: jpg, jpeg, png, bmp, webp. See LivePortrait Video Generation for details.

Failed to download input files.

Cause: Input file download failed.

Solution: Check that the file URL is accessible and the network is stable.

oss download error.

Cause: Input image download failed.

Solution: Check that the OSS link for the image is correct and accessible.

The image content does not comply with green network verification.

Cause: Image content violates compliance policies.

Solution: Replace the image with one that complies with Content Moderation policies.

read video error.

Cause: Failed to read the video.

Solution: Check if the video file is corrupted or if the format is unsupported.

the size of input image is too small or too large.

Cause: Input image dimensions are too small or too large.

Solution: Adjust the image dimensions to meet API requirements.

The request parameter is invalid, please check the request parameter.

Cause: The clothes_type parameter is non-compliant.

Solution: See OutfitAnyone - Image Segmentation for details.

The type or value of {parameter} is out of definition.

Cause: Parameter type or value does not meet requirements.

Solution: See LivePortrait Video Generation for details.

The request parameter is invalid, please check the request parameter.

Cause: Aspect ratio parameter is non-compliant.

Solution: Valid options are "1:1" or "3:4".

request timeout after 23 seconds.

Cause: No data was sent to the service for over 23 seconds. This error occurs when using speech recognition (Paraformer), and real-time speech synthesis (CosyVoice).

Solution: Check why no data was sent to the server for an extended period. If no message is sent to the server for more than 23 seconds, end the task promptly.

Please ensure input text is valid.

Cause: If you are using real-time speech synthesis (CosyVoice), this error usually occurs because no text was sent for synthesis. Possible causes include the following: missing parameter (no value assigned to the text parameter) or code exceptions (assignment to text failed).

Solution: Debug your code to ensure the text parameter is correctly assigned and sent.

Missing required parameter 'xxx'! Please follow the protocol!

Error examples:

• Missing required parameter 'payload.model'! Please follow the protocol!

• Missing required parameter 'payload.task_group'! Please follow the protocol!

Cause:

1. WebSocket event JSON format is incorrect (general cause).

   When calling models using the WebSocket protocol, the sent JSON format is invalid. Common issues include the following:

   • Incorrect JSON nesting—the parameters are not placed in the correct hierarchical level.

   • Parameter name misspelled—for example, task_group written as taskgroup.

   • Parameter value is empty or not properly assigned.

2. Calling call() method not re-invoked after stop() (specific scenario).

   This applies only to: DashScope Java SDK for Fun-ASR or Paraformer real-time speech recognition.

   After calling the stop() method to end recognition, you did not re-call the call() method before sending more data, causing an invalid session state and triggering this error.

   Difference between SDK users and WebSocket users: "Missing required parameter" is a protocol-level error returned by the server and appears only when using the WebSocket protocol directly. If you use the DashScope Java SDK Recognition class, the SDK intercepts this operation on the client side and throws an "Invalid state" exception ("Expected recognition state should be started, but current state is idle") instead of returning the "Missing required parameter" error.

Solution:

1. For cause 1, check the JSON format and parameters.

2. For cause 2, ensure that after each call to stop(), you re-call the call() method before the next recognition round. If using the DashScope Java SDK, watch for the client-side "Invalid state" exception; if using the WebSocket protocol directly, you will receive the "Missing required parameter" error.

[tts:]Engine return error code: 418

Cause: When using real-time speech synthesis (CosyVoice), the request parameter voice (timbre) is incorrect, or the model and voice versions do not match.

Solution:

1. Check voice parameter assignment:

   • If using default timbres, verify against the "voice parameter" section in the Python SDK.

   • If using cloned timbres, use the CosyVoice Voice Cloning/Design API to confirm the timbre status is "OK" and that the timbre belongs to the same account as the calling account.

2. Check version compatibility: v2 models can only use v2 timbres; v1 models can only use v1 timbres. Do not mix versions.

Request voice is invalid!

Cause: If you are using real-time speech synthesis (CosyVoice), this error usually occurs because no timbre was set.

Solution: Check that the voice parameter is assigned a value. If you are using the WebSocket API reference, configure the parameters in the correct JSON format as specified in the API documentation.

ref_images_url and obj_or_bg must be the same length.

Cause: When using the multi-image reference feature in Wanxiang - Video Editing (2.1), the array lengths of ref_images_url and obj_or_bg do not match.

Solution: Ensure the array lengths of ref_images_url and obj_or_bg are identical.

check input data style.

Cause: Input parameters do not meet 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 calling via code, do not set 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 (width and height) of the image passed to the Qwen VL model do not meet model requirements.

• The output image size exceeds the 10 MB limit.

Solution:

• Image dimensions must satisfy:

  • Both width and height are at least 10 pixels.

  • Aspect ratio must not exceed 200:1 or 1:200.

• Adjust the parameters for generated images.

Wrong Content-Type of multimodal url

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

Supported Content Types for Qwen VL models: image/bmp, image/icns, image/x-icon, image/jpeg, image/jp2, image/png, image/sgi, image/tiff, image/webp. See Image formats supported by Qwen VL models for details.

Solution:

Field required: image_url

Cause: Missing input parameter image_url.

Solution: Refer to Emoji Video Generation and pass the image_url parameter.

Field required: driven_id

Cause: Missing input parameter driven_id.

Solution: Refer to Emoji Video Generation and pass the driven_id parameter.

Invalid ext_bbox

Cause: The input ext_bbox parameter is invalid.

Solution: Refer to Emoji Video Generation and pass the correct ext_bbox.

Driven not exist: driven_id

Cause: The input driven_id does not exist.

Solution: Refer to Emoji Video Generation and pass the correct driven_id.

Text request limit violated, expected 1.

Cause: When calling the CosyVoice speech synthesis WebSocket API reference, you set enable_ssml to true and sent the continue-task instruction multiple times.

Solution: When enable_ssml is set to true, you can send the continue-task instruction only once.

SSML text is not supported at the moment!

Cause: When using CosyVoice speech synthesis, the current model or timbre does not support SSML, or SSML functionality is not properly enabled.

Solution: Troubleshoot according to Limits and Constraints.

[tts:]Engine return error code: 428

Cause: The CosyVoice speech synthesis instruction parameter is used incorrectly. Specific issues may include the following:

• Length exceeded: instruction cannot exceed 100 characters (Chinese characters count as 2 characters; others as 1).

• Format or language error: Only the following models support instruction, with different rules:

  • cosyvoice-v3.5-flash, cosyvoice-v3.5-plus: free-form instructions allowed (e.g., emotion, speed).

  • cosyvoice-v3-flash:

    • Must follow a fixed format.

    • Only Chinese instruction is supported.

    • Supported instruction varies by timbre. See CosyVoice Timbre List for details.

Solution:

• Check that the instruction character length does not exceed 100.

• Confirm that your model supports the instruction parameter.

• When using cosyvoice-v3-flash, ensure instruction is in Chinese and follows the fixed format for the corresponding timbre.

At least one of 'lyrics' or 'prompt' must be provided.

Cause: When using the Fun-Music model, the request did not include the lyrics or prompt parameter.

Solution: Include at least one of lyrics or prompt in the request.

Lyrics content is illegal and cannot be used for music generation.

Cause: When using the Fun-Music model, the lyrics content failed compliance checks, possibly containing infringing material.

Solution: Modify the lyrics to ensure they contain no infringing or non-compliant content, then retry.

400-invalid_request_error-invalid_value

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

Cause: When using the OpenAI compatible protocol, the seed parameter is not within the [0, 2³¹-1] range.

Solution: Set the seed parameter within the [0, 2³¹-1] range.

400-invalid_request_error

you must provide a model parameter.

Cause: The request did not include the model parameter.

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 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: During Qwen-TTS real-time speech synthesis, the selected timbre was generated using Qwen-TTS voice cloning, but the models used differ.

Solution: Check that the target_model parameter used during voice cloning matches the model parameter used during speech synthesis.

400-Arrearage

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

Cause: The Alibaba Cloud account associated with the API Key has an overdue payment, causing access denial.

Solution: Go to Expenses and Costs to check for overdue payments:

• No overdue payment: Confirm that the API Key belongs to the current account.

• Overdue payment: Recharge promptly. After recharging, system balance updates may be delayed; wait a moment before retrying.

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: Input or output contains suspected sensitive content blocked by content moderation.

Solution: Modify the input content and retry.

Input xxx data may contain inappropriate content.

Cause: Input data (such as prompts or images) may contain sensitive content.

Solution: Perform content compliance checks and modify the input before retrying.

400-APIConnectionError

Connection error.

Cause: Local 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 using speech recognition (Paraformer) for recorded file recognition, the file to be recognized failed to download.

Solution: Check that the audio file URL is accessible via the public network.

400-InvalidFile.AudioLengthError

Audio length must be between 1s and 300s.

Cause: Audio length does not meet requirements.

Solution: Ensure audio duration is within [1, 300] seconds.

Audio length must be between 1s and 180s.

Cause: Audio length does not meet requirements.

Solution: Ensure audio duration is within [1, 180] seconds.

400-InvalidFile.NoHuman

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

Cause: The input image contains no person or no face was detected.

Solution: Upload a single-person photo.

400-InvalidFile.BodyProportion

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

Cause: The person’s proportion in the uploaded image does not meet requirements.

Solution: Upload an image that meets the person proportion requirements.

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 in the uploaded image does not meet requirements (face must be visible with minimal head tilt).

Solution: Upload an image that meets requirements.

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

Cause: The facial orientation in the uploaded image does not meet the requirements (the face must not have a significant offset).

Solution: Ensure the face in the image has no tilt.

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

Cause: The facial pose in the uploaded image does not meet the requirements (the face must not have a significant offset).

Solution: Ensure the face in the image has no tilt.

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: Uploaded image size does not meet requirements.

Solution: Image resolution must not exceed 7000×7000 and must be at least 400×400.

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 224.

Cause: Uploaded image size does not meet requirements.

Solution: Image resolution must have a longest side under 4096 pixels and a shortest side over 224 pixels.

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: Uploaded image size does not meet requirements.

Solution: Image resolution must not exceed xxx×xxx and must be at least 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: Uploaded image size does not meet requirements.

Solution: Image aspect ratio must be less than xxx, and resolution must not exceed yyy×yyy.

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

Cause: Video resolution does not meet requirements.

Solution: Video side length must be between xxx and yyy.

400-InvalidFile.FPS

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

Cause: Video frame rate does not meet requirements.

Solution: 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: Uploaded image is too dark.

Solution: Ensure the 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: Uploaded image shows the person from behind.

Solution: Ensure the person faces the camera directly.

400-InvalidFile.FullFace

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

Cause: Facial pose in the uploaded image does not meet requirements (face must be visible).

Solution: Ensure the face in the image is complete and unobstructed.

400-InvalidFile.FaceNotMatch

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

Cause: Face matching between reference image and video failed.

Solution: See VideoRetalk Video Generation for details.

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: Choose a video clip that includes a person.

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: Choose a video where the person occupies a larger proportion in the first frame.

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 unclear.

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

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

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

Solution: Upload a single-person photo.

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

Cause: The input image contains an incomplete or missing human body.

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

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

Cause: The input image contains multiple people.

Solution: Upload a single-person photo.

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 full-body.

Solution: The full 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 person’s pose in the uploaded image does not meet requirements.

Solution: Upload an image that meets requirements: for profile photos, the head must be fully visible; for half-body photos, the area above the hips must be fully visible; or adjust the image 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 single person’s pose does not meet requirements.

Solution: Upload an image that meets requirements: shoulders and ankles must be visible, not from behind, not seated, and with minimal body tilt.

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: File size does not meet requirements.

Solution: Video files must be smaller than 200 MB; audio files must be smaller than 15 MB.

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

Cause: File size does not meet requirements.

Solution: Image files must be smaller than 5 MB.

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

Cause: File size does not meet requirements.

Solution: Video/audio/image files must be smaller than the specified MB limit.

400-InvalidFile.Duration

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

Cause: File duration does not meet requirements.

Solution: Video/audio file duration must be between xxx and yyy seconds.

400-InvalidFile.ImageSize

The size of image is beyond limit.

Cause: Image size exceeds the limit.

Solution: Image aspect ratio must not exceed 2, and the longest side must not exceed 4096.

400-InvalidFile.AspectRatio

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

Cause: File aspect ratio does not meet requirements.

Solution: Video file aspect ratio 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: File aspect ratio does not meet requirements.

Solution: Image file aspect ratio must be between 2.0 and 0.5.

400-InvalidFile.Openerror

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

Cause: File cannot be opened.

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

400-InvalidFile.Template.Content

Invalid template content.

Cause: Action template lacks permissions or content does not meet requirements.

Solution: Check 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: File format does not meet requirements.

Solution: Use supported formats: video—mp4, avi, mov; audio—mp3, wav, aac; image—jpg, jpeg, png, bmp, webp.

400-InvalidFile.MultiHuman

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

Cause: The input image contains multiple people.

Solution: Upload a single-person photo.

400-InvalidPerson

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

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

Solution: Upload a single-person photo.

400-InvalidParameter.DataInspection

Unable to download the media resource during the data inspection process. 

Cause: Downloading image or audio files timed out.

Solution: If calling from outside China, cross-border network instability may cause download timeouts. Store files in domestic OSS before calling the model. Alternatively, use the temporary storage space to upload files.

400-FlowNotPublished

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

Cause: The flow is not published.

Solution: Publish the flow and retry.

400-InvalidImage.ImageSize

The size of image is beyond limit.

Cause: Image size exceeds the limit.

Solution: Image aspect ratio must not exceed 2, and the longest side must not exceed 4096.

400-InvalidImage.NoHumanFace

No human face detected.

Cause: No face detected (only for asynchronous query interfaces of generation tasks).

Solution: Upload an image containing a clear human face.

400-InvalidImageResolution

The input image resolution is too large or small.

Cause: Input image resolution is too large or too small.

Solution: Image resolution must be at least 256×256 pixels and no more than 5760×3240 pixels.

400-InvalidImageFormat

The input image is in invalid format.

Cause: Image format does not meet requirements.

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

400-InvalidURL

Invalid URL provided in your request.

Cause: URL is invalid.

Solution: Use a valid URL.

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

Cause: Input URL is invalid or missing.

Solution: Provide a correct URL.

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

Cause: Input URL is invalid.

Solution: Ensure the URL is correct and points to an image file.

The input audio is longer than xxs.

Cause: Input audio file exceeds the maximum duration of xx seconds.

Solution: Trim the audio file to under xx seconds.

File size is larger than 15MB.

Cause: Input audio file exceeds the 15 MB limit.

Solution: Compress the audio file to under 15 MB.

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

Cause: Input audio format is non-compliant.

Solution: 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: Image is inaccessible or downloaded file format is unsupported.

Solution: Ensure 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: Image file format is unsupported.

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

400-InvalidURL.ConnectionRefused

Connection to xxx refused, please provide available URL.

Cause: Download was refused.

Solution: Provide an accessible URL.

400-InvalidURL.Timeout

Download xxx timeout, please check network connection.

Cause: Download timed out.

Solution: Check network connectivity.

400-BadRequestException

Invalid part type.

Cause: Applies only to Qwen-Long model conversation scenarios where the user uploaded a file type not yet supported by Qwen-Long.

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

400-BadRequest.EmptyInput

Required input parameter missing from request.

Cause: The request did not include the input parameter.

Solution: Add the input parameter to the request.

400-BadRequest.EmptyParameters

Required parameter "parameters" missing from request.

Cause: The request did not include the parameters parameter.

Solution: Add the parameters parameter to the request.

400-BadRequest.EmptyModel

Required parameter "model" missing from request.

Cause: The request did not provide the model parameter.

Solution: Add the model parameter to the request.

400-BadRequest.IllegalInput

The input parameter requires json format.

Cause: Input parameter format does not meet the API’s JSON requirements.

Solution: Check the input parameter format and ensure it is valid JSON.

400-BadRequest.InputDownloadFailed

Failed to download the input file: xxx.

Cause: Input file download failed, possibly due to timeout, download failure, or file size exceeding limits.

Solution: Troubleshoot based on the detailed error message xxx.

Failed to download the input file.

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

Solution: Check if the audio file can be downloaded normally. If it can, verify that the file size does not exceed the limit (10 MB).

400-BadRequest.UnsupportedFileFormat

File format unsupported.

Cause: When using CosyVoice voice cloning, the uploaded audio format does not meet model requirements.

Solution: Audio format must be WAV (16bit), MP3, or M4A. Note that file extension alone is not reliable—for example, a file with a .mp3 extension might actually be Opus. Use tools (such as ffprobe, mediainfo) or commands (like the Linux/macOS file command) to verify the actual audio encoding format.

Input file format is not supported.

Cause: Input file format is unsupported.

Solution: Use a supported file format.

400-BadRequest.TooLarge

Payload Too Large.

Cause: File size exceeds the limit.

Solution:

• When "purpose" is "file-extract", documents must not exceed 150 MB and images must not exceed 20 MB.

• When "purpose" is "batch", files must not exceed 500 MB. Split and upload files in batches.

400-BadRequest.ResourceNotExist

The Required resource not exist.

Cause:

• When calling update, query, or delete interfaces for CosyVoice voice cloning, the corresponding timbre does not exist.

400-Throttling.AllocationQuota

Your current quota is xxx

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

Solution: Delete some timbres.

Maximum voice storage limit exceeded, please delete existing voices.

Cause: When using Qwen-TTS voice cloning, you exceeded the main account’s timbre limit.

Solution: Delete some timbres.

400-InvalidGarment

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

Cause: Clothing image is missing.

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

400-InvalidSchema

Database schema is invalid for text2sql.

Cause: Database schema information was not provided.

Solution: Enter database schema information.

400-InvalidSchemaFormat

Database schema format is invalid for text2sql.

Cause: Input table information format is invalid.

Solution: Check and correct the table information format.

400-Audio.AudioShortError

valid audio too short!

Cause: Audio duration for CosyVoice voice cloning is too short.

Solution: Keep audio duration around 10–15 seconds. Ensure continuous speech with at least one segment over 5 seconds.

400-Audio.AudioSilentError

silent audio error.

Cause: CosyVoice voice cloning audio file is silent or non-silent segments are too short.

Solution: Keep audio duration around 10–15 seconds and include at least one continuous speech segment over 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: Image dimensions or file size do not meet requirements.

Solution: See Input Image Requirements.

400-FaqRuleBlocked

Input or output data is blocked by faq rule.

Cause: Triggered FAQ rule intervention module.

400-ClientDisconnect

Client disconnected before task finished!

Cause: The client disconnected before task completion. This error occurs when using speech synthesis or recognition services.

Solution: Check your code to avoid disconnecting from the server before task completion.

400-ServiceUnavailableError

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

Cause: Input content length is 0 or role is incorrect.

Solution: Ensure input content length is greater than 0 and parameter format (such as role) complies with API documentation.

400-IPInfringementSuspect

Input data is suspected of being involved in IP infringement.

Cause: Input data (such as prompts or images) is suspected of intellectual property infringement.

Solution: Perform content compliance checks and ensure input contains no infringing content.

400-UnsupportedOperation

The operation is unsupported on the referee object.

Cause: The associated object does not support this operation.

Solution: Verify that the operation object and operation type match.

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

Cause: Cannot delete the fine-tuning job because its status is already "succeeded," "failed," or "canceled."

Solution: Only jobs in specific states can be deleted; do not delete jobs in terminal states.

400-CustomRoleBlocked

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

Cause: Request or response content failed custom policy checks.

Solution: Check content or adjust custom policies.

400-Audio.PreprocessError

Audio preprocess error.

Cause: When using Qwen-TTS voice cloning, preprocessing of the audio to be cloned failed, possibly because: the text parameter content differs greatly from the audio transcript, valid speech is too short, or there is no sound.

Solution: Adjust the text parameter content. If ineffective, re-record the audio following the recording guidelines.

No segments meet minimum duration requirement

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

Solution: Re-record the audio following the recording guidelines.

400-BadRequest.VoiceNotFound

Voice '%s' not found.

Cause: When using Qwen-TTS voice cloning, the timbre specified in the delete interface call has already been deleted or does not exist.

Solution: Verify that the voice parameter is correct.

400-Audio.DecoderError

Decoder audio file failed.

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

Solution: Check if the audio file is corrupted and ensure it meets format requirements (WAV (16bit), MP3, or M4A for CosyVoice).

400-Audio.AudioRateError

File sample rate unsupported.

Cause: When using Qwen-TTS or CosyVoice voice cloning, the sample rate of the audio to be cloned does not meet requirements.

Solution: Sample rate must be at least 24,000 Hz.

400-Audio.DurationLimitError

Audio duration exceeds maximum allowed limit.

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

Solution: Audio must not exceed 60 seconds.

401-InvalidApiKey/invalid_api_key

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

Cause: API Key is incorrect.

Solution: Common causes and fixes:

• Reading wrong environment variable

  • Incorrect: api_key=os.getenv("sk-xxx") —the system tries to read an environment variable named sk-xxx, not use sk-xxx as the key.

  • Correct:

    • If environment variable is configured: Use api_key=os.getenv("DASHSCOPE_API_KEY");

      Ensure the DASHSCOPE_API_KEY environment variable is set before running.

    • If environment variable is not configured: Use api_key = "sk-xxx".

      This method is for debugging only; do not use in production.

• Typo: Alibaba Cloud Model Studio API Keys start with sk-. Confirm you did not mistakenly use another provider’s key and that copying did not include extra spaces or line breaks.

• Coding Plan exclusive API Key: Coding Plan provides an exclusive API Key (starting with sk-sp-) that must be used with the exclusive API endpoint (e.g., https://coding-intl.dashscope.aliyuncs.com/v1), unlike general API Keys. Confirm you updated both the API Key and Base URL. See Integrating AI Tools for configuration details.

• Region mismatch: API Key and Base URL belong to different regions—for example, using a China (Beijing) API Key with a Singapore Base URL. Confirm your API Key is from the Singapore region page or Beijing region page, or the US region page. Base URLs by region:

  Region	OpenAI compatible	DashScope
  Singapore	https://{WorkspaceId}.ap-southeast-1.maas.aliyuncs.com/compatible-mode/v1	https://{WorkspaceId}.ap-southeast-1.maas.aliyuncs.com/api/v1
  China (Beijing)	https://dashscope.aliyuncs.com/compatible-mode/v1	https://dashscope.aliyuncs.com/api/v1
  US (Virginia)	https://dashscope-us.aliyuncs.com/compatible-mode/v1	https://dashscope-us.aliyuncs.com/api/v1

  Important

  Model Studio has released workspace-specific domains for the China (Beijing) and Singapore regions. The new dedicated domains deliver superior performance and higher stability for inference requests. We recommend migrating to the new domains:

  • China (Beijing): from https://dashscope.aliyuncs.com to https://{WorkspaceId}.cn-beijing.maas.aliyuncs.com

  • Singapore: from https://dashscope-intl.aliyuncs.com to https://{WorkspaceId}.ap-southeast-1.maas.aliyuncs.com

  {WorkspaceId} is your workspace ID, which can be found on the Workspace Details page in the Model Studio console. The existing domain remains fully functional.

• Tool compatibility issue: Third-party tools are not properly adapted (e.g., latest Dify plugin instability causes errors; try installing an older Qwen plugin version; for older Cline, select OpenAI-Compatible instead of Alibaba Qwen as the API Provider).

If none apply, your API Key may have been deleted; reacquire and retry.

401-NOT AUTHORIZED

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

Cause:

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

• Or the request endpoint (service endpoint) is incorrect.

Solution:

• Confirm WorkspaceId is correct and the account is a workspace member before calling the API.

• China site users: use the China (Beijing) region endpoint; international site users: use the Singapore region endpoint. When using online debugging, confirm the service endpoint is correct.

  

401-invalid access token or token expired

invalid access token or token expired.

Possible cause: Token Plan mistakenly used Coding Plan or another plan’s Base URL.

Solution: Use the Token Plan exclusive Base URL:

• Anthropic compatible endpoint: https://token-plan.cn-beijing.maas.aliyuncs.com/apps/anthropic

• OpenAI compatible endpoint: https://token-plan.cn-beijing.maas.aliyuncs.com/compatible-mode/v1

403-AccessDenied/access_denied

Current user api does not support asynchronous calls.

Cause: The API does not support asynchronous calls.

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

current user api does not support synchronous calls.

Cause: The API does not support synchronous calls.

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

Invalid according to Policy: Policy expired.

Cause: The file upload credential expired when obtaining a temporary public URL.

Solution: Re-call the file upload credential API to generate a new credential.

Access denied.

Cause: No permission to access this model. The model may require approval, or its free quota is exhausted and it does not support pay-as-you-go (e.g., deepseek-r1-distill-llama-70b).

403-AccessDenied.Unpurchased

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

Cause: Alibaba Cloud Model Studio service is not activated.

Solution: Follow the steps below to activate the Alibaba Cloud Model Studio service.

1. Create an account: If you don’t have an Alibaba Cloud account, register first.

2. Select a region: Alibaba Cloud Model Studio supports multiple regions with different endpoints, models, and pricing. Go to the Model Studio console to choose a suitable region.

3. Complete identity verification: Use your Alibaba Cloud account to complete identity verification.

   • Choose individual users or upgrade to enterprise based on your situation, then click Verify Now.

   • For more information, see Identity Verification Overview.

4. Activate Alibaba Cloud Model Studio: Use your Alibaba Cloud account to go to Alibaba Cloud Model Studio, read and agree to the terms, and the service will activate automatically. If the service agreement does not appear, you are already activated.

403-Model.AccessDenied

Model access denied.

Cause: No permission to call the specified standard model.

Solution:

• Calling standard models: When using a sub-workspace API key to call standard models (e.g., qwen-plus), the sub-workspace must have calling permissions. See Model Calling Authorization.

• Calling custom models: After successful deployment, custom models can only be called with the API key of their workspace and do not require model calling authorization.

403-App.AccessDenied

App access denied.

Cause: No permission to access the application or model.

Solution:

• Confirm access permissions are granted for the workspace and RAM user.

• Verify the application is published.

• Double-check the APP ID and API KEY.

• If this is a Claude Code error, use the default workspace API Key.

• If all above are correct, refresh data, republish, and retry, or recreate the agent.

403-Workspace.AccessDenied

Workspace access denied.

Cause: No permission to access applications or models in the workspace.

Solution:

• If calling a sub-workspace model, see Sub-workspace Model Calling.

• Alternatively, use the main account’s API KEY, which has permissions for all workspaces.

403-Endpoint.AccessDenied

Workspace endpoint access denied.

Cause: You might be calling a deprecated model (e.g., historical snapshot versions like qwen-max-2025-01-25). After deprecation, the endpoint is no longer available, causing this error.

Solution:

• Go to Model Deprecation Mechanism to confirm if the model is deprecated.

• If deprecated, switch to its recommended replacement model.

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 1: New users exhaust free quota and make requests.

Solution: Complete sign-up before calling.

Cause 2: You enabled free quota exhaustion stop and made requests after free quota ran out.

Console free quota updates every minute (refresh manually).

Solution:

• To pay for calls, wait until the console shows free quota is exhausted, then disable the free quota exhaustion stop button.

• If using Coding Plan, this is usually a configuration error. Coding Plan requires a dedicated Base URL and API Key. See Coding Plan QuickStart for details.

404-ModelNotFound/model_not_found

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

Cause: The model does not support Batch calls, or the model name is misspelled.

Solution: See OpenAI compatible - Batch (file input) to confirm supported models and 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 does not exist or you haven’t activated Alibaba Cloud Model Studio.

Solution:

• Compare against model names in the Model List to verify your input (value of model parameter).

• Go to the Model Marketplace to activate model services.

404-model_not_supported

Unsupported model xxx for OpenAI compatibility mode.

Cause: The model does not support OpenAI compatible access.

Solution: Use the DashScope native method to call.

404-WorkSpaceNotFound

WorkSpace can not be found.

Cause: Workspace does not exist.

404-NotFound

Not found!

Cause:

• The resource to query/operate does not exist.

Solution:

• Check if the resource ID to query/operate is incorrect.

Request path not found.

Cause: When using the Fun-Music model, the service endpoint does not exist.

Solution: Check the API path for correctness and abnormal characters.

429-Throttling

Requests throttling triggered.

Cause: API calls triggered rate limiting.

Solution: Reduce call frequency or retry later.

Too many requests in route. Please try again later.

Cause: Too many requests triggered rate limiting.

Solution: Retry later.

429-Throttling.RateQuota/LimitRequests/limit_requests

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

Cause: Call frequency (RPS/RPM) triggered rate limiting.

Solution: See Rate Limiting to control 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: Call frequency spiked suddenly before hitting rate limits, triggering system stability protection.

Solution: Optimize client logic to use smooth request strategies (e.g., uniform scheduling, exponential backoff, or request queue buffering) to distribute requests evenly over time and avoid spikes.

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: Token consumption per second or minute (TPS/TPM) triggered rate limiting.

Solution: See the Rate Limiting documentation for model limits and adjust calling strategy. If default quotas are insufficient, request a temporary TPM increase in the Model Studio console under Rate Limit Increase.

See FAQ to avoid triggering rate limits.

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 now; retry later.

Free allocated quota exceeded.

Cause: Free quota expired or exhausted, and the model does not support pay-as-you-go.

Solution: Replace with another model—for example, if Qwen Audio model quota is exhausted, use the non-real-time (Qwen-Omni) model.

Maximum voice-clone voice limit exceeded.

Cause: When using Qwen-TTS voice cloning, you exceeded the main account’s timbre limit.

Solution: Delete some timbres.

429-CommodityNotPurchased

Commodity has not purchased yet.

Cause: Workspace subscription not purchased.

Solution: Purchase workspace service first.

429-PrepaidBillOverdue

The prepaid bill is overdue.

Cause: Workspace prepaid bill expired.

429-PostpaidBillOverdue

The postpaid bill is overdue.

Cause: Model inference service expired.

430-Audio.DecoderError

Decoder audio file failed.

Cause: CosyVoice voice cloning audio file decoding failed.

Solution: Use tools (such as ffprobe, mediainfo) or commands (like the Linux/macOS file command) to verify the actual audio encoding format.

430-Audio.FileSizeExceed

File too large

Cause: CosyVoice voice cloning audio file size exceeds limit.

Solution: Audio file for voice cloning must be under 10 MB.

430-Audio.AudioRateError

File sample rate unsupported

Cause: CosyVoice voice cloning audio file sample rate is unsupported.

Solution: Set sample rate to 16 kHz or higher.

430-Audio.AudioSilentError

Silent file unsupported.

Cause:CosyVoice voice cloning audio file is silent or non-silent segments are too short.

Solution: Keep audio duration around 10–15 seconds and include at least one continuous speech segment over 5 seconds.

500-InternalError/internal_error

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

Cause: Internal error.

Solution:

• If using the (Qwen-Omni) model, use streaming output.

• If using CosyVoice voice cloning, possible causes:

  • Audio file is non-standard—for example, poor audio quality, noise, or volume fluctuations. See Recording Guidelines and retry.

  • Recording file URL is inaccessible. Follow instructions in CosyVoice Voice Cloning/Design API and retry.

  • Recording file is too long. Use recordings around 10–15 seconds with at least one continuous speech segment over 5 seconds.

Internal server error!

Cause: Internal algorithm error.

Solution: Retry later.

audio preprocess server error

Using CosyVoice voice cloning:

• Cause: Audio file is non-standard—for example, poor audio quality, noise, or volume fluctuations.

  Solution: See Recording Guidelines and retry.

• Cause: Recording file URL is inaccessible,

  Solution: Follow instructions in CosyVoice Voice Cloning/Design API and retry.

• Cause: Recording file is too long.

  Solution: Use recordings around 10–15 seconds with at least one continuous speech segment over 5 seconds.

request asr failed

Cause: When using CosyVoice voice cloning, the audio file is non-standard—lacking valid speech or having unclear audio with heavy noise.

Solution: See Recording Guidelines and retry.

Remote cancelled grpc stream.

Cause: The timbre used in speech synthesis does not exist.

Solution: Check the voice parameter and ensure a valid timbre name is specified. See Real-time Speech Synthesis (CosyVoice) for available timbres.

500-InternalError.FileUpload

oss upload error.

Cause: File upload failed.

Solution: Check OSS configuration and network.

500-InternalError.Upload

Failed to upload result.

Cause: Result upload failed.

Solution: Check storage configuration or retry later.

500-InternalError.Algo

inference internal error.

Cause: Service exception.

Solution: Retry first to rule out transient issues.

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

Cause: Model-generated JSON data is invalid, preventing tool calls.

Solution: Switch to the latest model or optimize prompts and retry.

Missing Content-Length of multimodal url.

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

Solution: If unresolved, try another image link.

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

Cause: Model service backend server error.

Solution: Retry later.

An internal error has occured during algorithm execution.

Cause: Algorithm runtime error.

Solution: Retry later.

Inference error: Inference error.

Cause: Inference error.

Solution: Check if the image file is corrupted or if the person image quality is sufficient (must contain a complete, clear face).

Role must be in [user, assistant]

Cause: When using the Qwen-MT model, the messages array contains non-user role messages.

Solution: Ensure the messages array contains only one element, which must be a user message (User Message).

Embedding_pipeline_Error: xxx

Cause: Image or video preprocessing error.

Solution: Confirm uploaded images/videos and request code meet requirements, then retry.

Receive batching backend response failed!

Cause: Internal service error.

Solution: Retry later.

[music]Receive batching backend response failed!

Cause: When using the Fun-Music model, service concurrency limits were exceeded.

Solution: Reduce concurrent requests and retry.

Other kinds of server error.

Cause: When using the Fun-Music model, unknown internal system exception.

Solution: Provide the request ID to technical staff for troubleshooting.

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: Internal algorithm error.

Solution: Retry later.

list index out of range

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

Solution: Adjust the messages array order to ensure 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: Asynchronous task did not return results within 3 hours, causing timeout.

Solution: Check task execution or contact support.

500-SystemError

An system error has occured, please try again later.

Cause: System error.

Solution: Retry later.

500-ModelServiceFailed

Failed to request model service.

Cause: Model service call failed.

Solution: Retry 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:

• Request timeout when calling large models; timeout error occurs after 300 seconds.

• When using speech recognition (Paraformer), no audio sent to server for a long time or prolonged silence.

• When calling image generation or editing models, processing time exceeds limits due to large image size or high complexity.

Solution:

• Use streaming output. See Streaming Output.

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

• When calling image models, try lowering resolution, simplifying edits, or retrying later.

500-ResponseTimeout

Response stream timeout

Cause: When using the Fun-Music model, internal execution timed out.

Solution: Retry the call.

500-InvokePluginFailed

Failed to invoke plugin.

Cause: Plugin invocation failed.

Solution: Check plugin configuration and availability.

500-AppProcessFailed

Failed to proceed application request.

Cause: Application flow processing failed.

Solution: Check application configuration and flow nodes.

500-RewriteFailed

Failed to rewrite content for prompt.

Cause: Large model call for prompt rewriting failed.

Solution: Retry later.

500-RetrivalFailed

Failed to retrieve data from documents.

Cause: Document retrieval failed.

Solution: Check document index and retrieval configuration.

500/503-ModelServingError

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

Cause: Network resources are saturated; your request cannot be processed temporarily.

Solution: Try again later.

503-ModelUnavailable

Model is unavailable, please try again later.

Cause: Model temporarily unavailable.

Solution: Retry 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: API Key not provided when using DashScope SDK.

Solution: See Configuring API Key as Environment Variable for details.

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: API Key not passed when using OpenAI SDK.

Solution:

• Pass API Key via environment variable (recommended)

  Set DASHSCOPE_API_KEY as an environment variable (see Configuring API Key as Environment Variable), then initialize the client by reading it with os.getenv:

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

• Hardcode API Key (testing only)

  Directly pass the API Key to the api_key parameter:

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

  Note: This poses a security risk; do not use in production.

Bad Request for url: xxx

Cause: When using Python requests library, adding response.raise_for_status() causes errors without returning specific server error content.

Solution: Use print(response.json()) to view server response.

Cannot resolve symbol 'ttsv2'

Cause: If using real-time speech synthesis (CosyVoice), this issue arises from an outdated DashScope SDK version.

Solution: Install the latest DashScope SDK.

NetworkError

NoApiKeyException: Can not find api-key.

Cause: Environment variable configuration did not take effect.

Solution: Restart your client or IDE and retry. See FAQ for more scenarios.

ConnectException: Failed to connect to dashscope.aliyuncs.com

Cause: Local network environment is abnormal.

Solution: Check local network—for example, HTTPS access blocked by certificate issues or incorrect firewall settings. Try a different network or server.

InputRequiredException: Parameter invalid: text is null

Cause: When using real-time speech synthesis (CosyVoice), no text was sent for synthesis.

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

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

Cause: Current DashScope SDK version is outdated.

Solution: Install the latest 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 against OpenAI compatible - Batch (file input).

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, each request ID must be unique.

Solution: Check your input file against OpenAI compatible - Batch (file input) to ensure all request IDs are unique.

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

Cause: File upload failed because the Alibaba Cloud Model Studio storage space under your account is full or nearly full.

Solution: Delete unnecessary files via the OpenAI compatible - File interface to free space. Current storage supports up to 10,000 files totaling 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 using streaming speech recognition (Paraformer), the service returned recognition results that are too large.

Solution: Send audio in segments—about 100 ms per segment, with data size 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 using speech synthesis (CosyVoice), the WebSocket connection could not be established within 5 seconds.

Solution: Check local network, firewall settings, or try a different network or server.

unsupported audio format:xxx

Cause: When using CosyVoice voice cloning, the uploaded audio format does not meet model requirements.

Solution: Audio format must be WAV (16bit), MP3, or M4A. Do not rely solely on file extensions; use tools (such as ffprobe, mediainfo) or commands to verify the actual encoding format.

internal unknown error

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

Solution: Audio format must be WAV (16bit), MP3, or M4A. Use tools to verify the actual encoding format.

Invalid backend response received (missing status name)

Cause: When using the Paraformer recorded file recognition RESTful API, request parameters are misspelled.

Solution: Check code against API documentation.

NO_INPUT_AUDIO_ERROR

Cause: No valid speech detected.

Solution: If using Paraformer real-time speech recognition, troubleshoot as follows:

1. Check for audio input.

2. Check audio format (supported: pcm, wav, mp3, opus, speex, aac, amr).

SUCCESS_WITH_NO_VALID_FRAGMENT

Cause: If using Paraformer recorded file recognition, the recognition result query succeeded, but the VAD module detected no valid speech.

Solution: Check if the recording contains valid speech. If it’s all silence, no recognition result is normal.

ASR_RESPONSE_HAVE_NO_WORDS

Cause: If using Paraformer recorded file recognition, the recognition result query succeeded, but the final result is empty.

Solution: Check if the recording contains valid speech, or if valid speech consists only of fillers and the disfluency_removal_enabled parameter filtered them out.

FILE_DOWNLOAD_FAILED

Cause: If using Paraformer recorded file recognition, the file to be recognized failed to download.

Solution: Check if the recording file path is correct and accessible via the public network.

FILE_CHECK_FAILED

Cause: If using Paraformer recorded file recognition, file format is incorrect.

Solution: Check if the recording file is single-track/dual-track WAV or MP3 format.

FILE_TOO_LARGE

Cause: If using Paraformer recorded file recognition, the file to be recognized is too large.

Solution: Check if the recording file exceeds 2 GB; if so, split it.

FILE_NORMALIZE_FAILED

Cause: If using Paraformer recorded file recognition, file normalization failed.

Solution: Check if the recording file is corrupted or playable.

FILE_PARSE_FAILED

Cause: If using Paraformer recorded file recognition, file parsing failed.

Solution: Check if the recording file is corrupted or playable.

MKV_PARSE_FAILED

Cause: If using Paraformer recorded file recognition, MKV parsing failed.

Solution: Check if the recording file is corrupted or playable.

FILE_TRANS_TASK_EXPIRED

Cause: If using Paraformer recorded file recognition, the recognition task expired.

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

REQUEST_INVALID_FILE_URL_VALUE

Cause: If using Paraformer recorded file recognition, the file_link parameter is invalid.

Solution: Confirm the file_url parameter format is correct.

CONTENT_LENGTH_CHECK_FAILED

Cause: If using Paraformer recorded file recognition, content-length check failed.

Solution: Check that the HTTP response’s content-length matches the actual file size when downloading the recording.

FILE_404_NOT_FOUND

Cause: If using Paraformer recorded file recognition, the file to download does not exist.

Solution: Check if the file URL is correct.

FILE_403_FORBIDDEN

Cause: If using Paraformer recorded file recognition, no permission to download the recording.

Solution: Check file access permissions.

FILE_SERVER_ERROR

Cause: If using Paraformer recorded file recognition, the file server is unavailable.

Solution: Retry later or check file server status.

AUDIO_DURATION_TOO_LONG

Cause: If using Paraformer recorded file recognition, the file duration exceeds 12 hours.

Solution: Split the audio into segments and submit multiple recognition tasks. Use tools like FFmpeg to split.

DECODE_ERROR

Cause: If using Paraformer recorded file recognition, audio file info detection failed.

Solution: Confirm the download link points to a supported audio format.

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

Cause: When using Qwen-TTS real-time speech synthesis with model qwen-tts-vc-realtime-2025-08-20, you used a default timbre. This model supports only cloned timbres.

Solution: Use a timbre generated via voice cloning, not a default timbre.

NO_VALID_AUDIO_ERROR

Cause: When using speech recognition (Paraformer), the audio to be recognized is invalid.

Solution: Check audio format, sample rate, etc., to ensure requirements are met.

InvalidParameter: task can not be null

Cause: When using CosyVoice speech synthesis WebSocket API, the run-task or finish-task payload lacks an input field, or the continue-task payload lacks input.text.

Solution:

1. Check run-task command: ensure payload contains "input": {} (empty object); do not omit input.

2. Check continue-task command: ensure payload.input contains a non-empty text field.

3. Check finish-task command: ensure payload contains "input": {}.

200- BailianGateway.Workspace.NotAuthorised

Cause: This error may occur due to: (1) special characters or non-standard formats in the access URL causing workspace authorization validation to fail; (2) RAM sub-account operating a workspace without permissions.

Solution: (1) Revisit the Model Studio console homepage and navigate to the target page; (2) the main account or a RAM account with administrative permission must grant the sub-account access to the corresponding workspace.

Coding Plan

Connection error

Cause: Base URL typo or network issue.

Solution: Check Base URL spelling and network connection.

hour allocated quota exceeded

Cause: 5-hour request quota exhausted.

Solution: Quota resets automatically after 5 hours.

week allocated quota exceeded

Cause: Weekly request quota exhausted.

Solution: Quota resets at 00:00:00 (UTC+8) every Monday.

month allocated quota exceeded

Cause: Monthly request quota exhausted.

Solution: Quota resets at 00:00:00 (UTC+8) on the subscription day each month.

concurrency allocated quota exceeded

Cause: Current concurrent requests exceed the platform’s dynamically allocated limit.

Solution: Retry shortly. The platform dynamically adjusts concurrency limits based on overall load; peak times may trigger this restriction.

usage allocated quota exceeded. please try again later.

Cause: Beyond call count limits, Coding Plan also evaluates short-term resource consumption. High short-term consumption triggers temporary rate limiting.

Solution: Usually recovers within an hour. Split large tasks into smaller ones and submit over time.