Baby vs LLM: Agent evaluation under operational disguise ( with source code )

Results are subject to change as I continue to complete it for the rest of the models. A few days...