S1M0N38-hn

3 ms·

1. 1.▲
   by S1M0N38-hn 8mo ago

   Im pretty open to feedback and contribution (also regarding the default strategy). So feel free to open Issues on GH. However I'd like to collect a bunch of them (including bugs) before re-running the whole benchmark (balatrobench v2).
2. 2.▲
   by S1M0N38-hn 8mo ago

   not really. I've downloaded balatro. I saw that it was moddable. I wrote a mod API to interact programmatically. I was just curious if, from text only game state representation, a LLM would be able to make some decent play. the benchma
3. 3.▲
   by S1M0N38-hn 8mo ago

   Hi, BalatroBench creator here. Yeah, Google models perform well (I guess the long context + world knowledge capabilities). Opus 4.6 looks good on preliminary results (on par with Gemini 3 Pro). I'll add more models and report soon. Tbh
4. 4.▲
   BalatroBench – Benchmarking LLMs' Strategic Performance Through Games(balatrobench.com)

   3 points by S1M0N38-hn 8mo ago|0 comments