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

Reasoning GYM

Benchmarks

Task NameDataset NameSOTA ResultTrend
ReasoningReasoning-Gym (test)
Accuracy70.1
28
ReasoningReasoning-Gym (train)
Accuracy63.3
28
General ReasoningReasoning GYM
Accuracy (Reasoning GYM)8.8
10
Logical ReasoningReasoning Gym Color Cube
Accuracy (*)32
8
Logical ReasoningReasoning Gym Leg Counting
Accuracy50.67
8
Logical ReasoningReasoning Gym Zebra Puzzles
Accuracy (*)39.33
8
ReasoningReasoning Gym Puzzle24
Success Rate60.67
8
ReasoningReasoning Gym Word Sorting
Success Rate84.33
8
ReasoningReasoning Gym Spell Backward
Success Rate97
8
ReasoningReasoning Gym Manipulate Matrix
Success Rate8.33
8
ReasoningReasoning Gym ARC-1D
Success Rate30.67
8
Showing 11 of 11 rows