I’ve been using the Gemini API in both Python and .NET. The issue I’m facing is that, when using the same prompt, the response in Python is accurate and as expected. For example, when processing a clear image in Python, the output is correct with an accuracy of 0.95. However, using the same image and prompt in .NET, the output is incorrect, and the accuracy drops to 0.2.his works well in Python, the same prompt in .NET does not yield consistent results. As well as on same prompt at different hit recieve different accuracy of the same image.