This is a rather unconventional yet insightful method for identifying LLMs. When you feed the 4th image into a multimodal LLM and prompt it with: "Describe this image in English, without using any tools.", different models yield distinctly different results. Currently, the only
These were converted using vectomancy.xuepoo.xyz. It is a program that converts images into mathematical curves. Feel free to use them!
The following images are all generated by gpt-image 2.
I switched to GPT-5.6 Sol for its strong balance of reasoning, speed, and coding quality.
I casually asked Claude a question, and he said there were security risks.
Wait! Don't Go Yet 🚀
Get our FREE eBook "10 Programming Tips That Changed Everything" when you subscribe to our newsletter!