Instruction Following
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
32.5Win-Loss Differential (%)
2
Feb 26, 2026
77.99Original Success Ratio
1
Jun 18, 2026
84.34Win Rate
1
May 14, 2026
3.25Judge 1 Score
1
Apr 24, 2026
63.9HATified SFT Score
1
Mar 18, 2026
63.3HATified SFT Score
1
Mar 18, 2026
66.1Win Rate
1
Mar 18, 2026
61.1Winrate
1
Mar 18, 2026
97HALT Wins
1
Feb 26, 2026
54HALT Wins
1
Feb 26, 2026
46HALT Wins
1
Feb 26, 2026
49HALT Wins
1
Feb 26, 2026
—
—Score
0
Feb 26, 2026
—
—Score
0
Feb 26, 2026
—
—Accuracy
0
Feb 26, 2026