Output error - Code test

Google AI Studio
1

I always get this error when trying to test if the model can code a simple snake game. It fails every time no matter what Gemini 1.5 model I use. I tried this same prompt for a few days straight and there hasn’t been any solution. I have tried the “continue” method, and told it to only do 10 lines at a time but it doesn’t work. To recap I have tried multiple variations on the prompt, but it consistently fails. It also fails when I use the main Gemini advanced interface.

It does not fail however when I choose Gemini 1.0, it’s just that it doesn’t do a very good job at the script. It also seems to work when I do the API on different platforms but that defeats the purpose of having a single platform to use. I pay for a few different AI platforms and would like to just pick one. I would like to use Gemini because I am an Android user. Have you guys seen this before?

Screenshot 2024-09-05 120006
Screenshot 2024-09-05 1200061432×587 60.4 KB

2

Welcome to the forum. There’s a typo in the prompt you show (game → came) which might be affecting the large language model output. The following prompt works fine:

Using Python, code a snake game for model view controller. Provide only the code for the model.

To find out what the problem is, click or tap on the orange triangle to find out what triggered the output block.

Hope that helps.

3

Most of the time I use game not came haha. I tried your prompt and while it doesn’t error out, it also is no where near successfully functionally (the game is not made). I have clicked on the Orange Triangle and it outputs the same error every time (which is shown in the previous image). Most LLM Platforms I use it has no issue with this basic prompt. Including Gemini1.5 pro on other platforms such as: Not Diamond. It is weird that it only fails on Google products.