Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

MTG

Benchmarks

Task NameDataset NameSOTA ResultTrend
Code GenerationMTG (test)
BLEU61.4
9
Music CaptioningMTG-eval-Cap 1.0 (test)
BLEU Score0.278
5
Story CompletionMTG Story Completion (test)
ROUGE-L (ES)31.2
3
MTG GameplayMTG Mono-Red Aggro in-distribution opponent pool
Win Rate70.6
2
MTG GameplayMTG Domain Ramp (in-distribution opponent pool)
Win Rate20.6
2
MTG GameplayMTG Dimir Midrange (in-distribution opponent pool)
Win Rate52.2
2
MTG GameplayMTG Boros Convoke (in-distribution opponent pool)
Win Rate57.2
2
MTG GameplayMTG Azorius Control (in-distribution opponent pool)
Win Rate25.6
2
Showing 8 of 8 rows