Troubleshooting of Google Gemini

This article "Troubleshooting of Google Gemini" is still being written. If there are any incomplete parts, you are welcome to directly edit them. 這篇文章「Troubleshooting of Google Gemini」內容還在撰寫中，如果有不完整的部分，歡迎你直接動手修改。

Troubleshooting of Google Gemini

How to fix "Request contains an invalid argument"

Error message

無法提交提示
錯誤訊息：「Request contains an invalid argument.」

狀態：400 錯誤代碼：3

要求 ID：xxx

Model Affected: gemini-1.5-pro-preview-0409 or gemini-1.5-flash-preview-0514

Workaround solution:

• If you upload a video file for API analysis, you might consider shortening the video duration, for example, using ffmpeg to reduce an hour-long video to half its length. This will help reduce the input token length.

How to fix: The File XXX is not in an ACTIVE state and usage is not allowed

After upload the video to Gemini API

Error message

Processing video...
$response is: Array
(
    [error] => Array
        (
            [code] => 400
            [message] => The File XXX is not in an ACTIVE state and usage is not allowed.
            [status] => FAILED_PRECONDITION
        )

)

Solution

• "You need to check its status and wait until it becomes ACTIVE" python - Google GenerativeAI unable to load video files - Stack Overflow

How to fix: Invalid or unsupported file uri

After upload the video, when I tried to analysis the video by the Gemini API

Error message

Uncaught Exception: API request failed with HTTP code 400
Response: {
  "error": {
    "code": 400,
    "message": "Invalid or unsupported file uri: ZGFkYjY0Njg0MDIyYmM3ZDc2YmIwY2Q5YzI4ZTllYzBhN2E1OGMzNDk2YWVkMmVlZWY0NDhlY2FhZGVjZTM1Ng==",
    "status": "INVALID_ARGUMENT"
  }
}

Solution

• Please verify the format of your file URL. It should follow this pattern:

`https://generativelanguage.googleapis.com/v1beta/files/XXXXX`

How to fix: Invalid JSON payload received. Unknown name "responseSchema" at 'generation_config'

After upload the video, when I tried to analysis the video by the Gemini API

Error message

{
  "error": {
    "code": 400,
    "message": "Invalid JSON payload received. Unknown name \"responseSchema\" at 'generation_config': Proto field is not repeating, cannot start list.",
    "status": "INVALID_ARGUMENT",
    "details": [
      {
        "@type": "type.googleapis.com/google.rpc.BadRequest",
        "fieldViolations": [
          {
            "field": "generation_config",
            "description": "Invalid JSON payload received. Unknown name \"responseSchema\" at 'generation_config': Proto field is not repeating, cannot start list."
          }
        ]
      }
    ]
  }
}

Solution

• The parameter "responseSchema" could not be an empty array

Resolving "Model not found for API version v1beta" Error with Gemini API

Error Details

{
  "error" : {
    "code" : 404,
    "message" : "models/gemini-2.0-flash-exp-image is not found for API version v1beta, or is not supported for generateContent. Call ListModels to see the list of available models and their supported methods.",
    "status" : "NOT_FOUND"
  }
}

Problematic Command

curl "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash-exp-image:generateContent" \
  -H 'Content-Type: application/json' \
  -H 'X-goog-api-key: GEMINI_API_KEY' \
  -X POST \
  -d '{
    "contents": [
      {
        "parts": [
          {
            "text": "Generate an image of a cute baby turtle in a 3d digital art style"
          }
        ]
      }
    ]
  }'

Affected Model: gemini-2.0-flash-exp-image-generation

Troubleshooting Steps:

# Set your API key
API_KEY="YOUR_API_KEY_HERE"
 
echo "Retrieving available Gemini models..."
curl -s "https://generativelanguage.googleapis.com/v1beta/models" \
  -H "X-goog-api-key: $API_KEY" | jq -r '.models[] | select(.name | contains("gemini")) | {name: .name, supportedGenerationMethods: .supportedGenerationMethods}'

Expected Output

{
  "name": "models/gemini-2.0-flash-exp-image-generation",
  "supportedGenerationMethods": [
    "generateContent",
    "countTokens",
    "bidiGenerateContent"
  ]
}

Issue Identified: Incorrect model name used. The valid model name is "models/gemini-2.0-flash-exp-image-generation", not "models/gemini-2.0-flash-exp-image"

Fixed Command

curl "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.0-flash-exp-image-generation:generateContent" \
  -H 'Content-Type: application/json' \
  -H 'X-goog-api-key: GEMINI_API_KEY' \
  -X POST \
  -d '{
    "contents": [
      {
        "parts": [
          {
            "text": "Generate an image of a cute baby turtle in a 3d digital art style"
          }
        ]
      }
    ]
  }'

How to resolve: Gemini Text-to-Image: 16:9 Aspect Ratio Not Working, Always Generates 1:1 Images

Question:

Has anyone experienced issues with Gemini's text-to-image function where you specify 16:9 aspect ratio but keep getting 1:1 square images?

Problematic prompt for Models: 2.5 Pro & 2.5 Flash

Generate a cute young turtle, 3D style
Image size 16:9

Generate a cute young turtle, 3D style
landscape orientation, wide format

Answer:

A Google Product Expert acknowledged this issue in official forums, describing it as a "known limitation." However, no timeline has been provided for when the functionality might be restored.

I've actually tested both Gemini 2.5 Pro and 2.5 Flash models, specifying image dimensions in the prompt (Image size: 16:9) or generating landscape orientation (landscape orientation, wide format), and both have the same problem. If you can make API calls, you might try using the API to call other models.

If you want an online solution that works, you can try Google Whisk directly: https://labs.google/fx/tools/whisk/ The interface allows you to directly set three aspect ratios: square 1:1, portrait 9:16, and landscape 16:9.

Further reading

• 疑難排解指南  |  Google AI for Developers  |  Google for Developers