I built Brewhaven, a MTG harness that tests the models on their ability to not only play the game, but also on diplomacy. So far only the recent models by Anthropic have been satisfactory on both domains. Unfortunately, they cost an arm and a leg so nowhere near financially reasonable to play right now.
I built Brewhaven, a MTG harness that tests the models on their ability to not only play the game, but also on diplomacy. So far only the recent models by Anthropic have been satisfactory on both domains. Unfortunately, they cost an arm and a leg so nowhere near financially reasonable to play right now.