EQ-Bench
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Polish EQ-Bench
78.07Overall Score
106
EQ-Bench Long-form writing (leaderboard)
0.798Score
23
EQ-Bench 3
0.896Score
20
EQ-Bench
86.4Overall Score
19
EQ-Bench Longform (Creative Writing) (32 prompts)
1,578Elo Rating
17
EQ-Bench Longform (ID prompts) (45 prompts)
1,911Elo Score
16
EQ-Bench Creative (OOD)
1,911Elo Rating
14
EQ-Bench Creative
84.6S Score
14
EQ-Bench3
83.3Overall Score
12
EQ-Bench Creative Writing v3
83.8EQ-Bench
6
EQ-Bench Longform (Far OOD 8–12k)
54.3EQ-Bench Score
1
EQ-Bench Longform Near OOD 4–8k
57.5EQ-Bench Score
1
EQ-Bench Longform (ID 1–4k)
63.1EQ-Bench Score
1
EQ-Bench Longform Aggregate
59.8EQ-Bench Score
1
EQ-Bench 3
—Primary metric
0