Instruction Following
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
6.31Score
14
May 29, 2026
79.91Win Rate
14
Feb 26, 2026
96.2Success Rate
14
Feb 26, 2026
95.8Normalized performance
13
Jun 16, 2026
82.92Strict Accuracy
13
Mar 31, 2026
76.81Strict Accuracy
13
Mar 31, 2026
79Strict Accuracy
13
Mar 31, 2026
62.81Strict Accuracy
13
Mar 31, 2026
93.2Overall Score
13
Mar 20, 2026
8.24Score
13
Feb 26, 2026
95.3AlpacaEval Score
13
Feb 26, 2026
85.1ROCG
12
Jul 7, 2026
94.3Accuracy (%)
12
Jun 16, 2026
99Instruction Following Accuracy
12
May 12, 2026
46.12PASS@1
12
Apr 21, 2026
17.269LC Win Rate
12
Apr 7, 2026
17.314LC Win Rate
12
Apr 7, 2026
61.9Win Rate
12
Mar 20, 2026
32.7Win Rate
12
Mar 20, 2026
3.85PPL
12
Feb 26, 2026
87.64Score
12
Mar 4, 2026
29.64LC Win Rate
12
Feb 26, 2026
53.37LC Win Rate
12
Feb 26, 2026
49.79pass@1
12
Feb 26, 2026
70.9Chinese Win Rate
12
Feb 26, 2026