>>717965068
Gemini is an LLM, but ChatGPT (also an LLM) beat it without any cheats. Just a very strong harness.

If you want to see a different type of AI that beat it without any "hints", then reinforcement learning has done it, and there's a few different projects, but I know of one that has beaten the game, and there might be more.

https://drubinstein.github.io/pokerl/

And here's a video to better visualize how this AI differs from LLMs. It's a bit old, and I'm not sure if this guy has made progress, but it's a good walkthrough on how everything works.

https://youtu.be/DcYLT37ImBY?si=QcB-sU46QXXRxc_E