LLM Comparison Test

Run the same prompt against 6 models in parallel. Lock 🔒 results you like, rotate 🔄 the rest.