express – Google Generative AI Node.js SDK – ApiError 500: Internal error encountered while calling generateContent()

I am integrating Google Generative AI (Gemini) Node.js SDK in my project.

I want to send audio input (webm chunks) to the model using generateContent().

Error processing chunk: ApiError:
{ “error”: { “code”: 500, “message”: “Internal error encountered.”, “status”:”INTERNAL”}}
at throwErrorIfNotOK (C:\Users\HP\Desktop\revolt-bot\node_modules@google\genai\dist\node\index.cjs:13977:30)
at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
at async C:\Users\HP\Desktop\revolt-bot\node_modules@google\genai\dist\node\index.cjs:13753:13
at async Models.generateContent (C:\Users\HP\Desktop\revolt-bot\node_modules@google\genai\dist\node\index.cjs:15021:24)
at async WebSocket. (C:\Users\HP\Desktop\revolt-bot\server.js:38:38) {status: 500}

Here is the part of my code where I’m calling the API:

while (retries <= maxRetries) {
  try {
    const response = await genAI.models.generateContent({
      model: "gemini-2.5-flash",
      contents: [
        {
          role: "user",
          parts: [
            {
              inlineData: {
                mimeType: "audio/webm",
                data: chunk.toString("base64"),
              },
            },
          ],
        },
      ],
    });

    console.log(response.response.text());
    break;
  } catch (err) {
    console.error("Error processing chunk:", err);
    retries++;
  }
}

• I have verified my API key and permissions.
• I have tried “gemini-2.5-flash” and “gemini-1.5-pro”.
• I have confirmed my audio chunks are valid WebM/Opus.
• I have retried with different base64 encoding methods.
• I updated the @google/genai package.