The game's core loop is being developed, with a barebones "benchmark mode" where players pit their trained pet detectives from their roster against detectives of other players. A battle for logical supremacy ensues and the winning detective gets evolution points. The images in the benchmark also have been tested against LLMs like GPT, LLAVA, Gemini, and their accuracy is represented as a percentile similar to IQ.
With enough training the Pet Detectives can even surpass these industry standard LLMs on specific domains!
The Benchmark mode will serve as a "Boss Battle" for the dataset, with several players pitting their detectives against each other for progression!