Is anyone giving out tokens for benchmarking LLMs?

I'm thinking of benchmarking different models performance playing Nethack and learning the game under different constraints.