Instruction Following
Benchmarks
DatasetSOTA methodMetricTrendResultsLast updated
35Rouge-L
72
Feb 26, 2026
82IFBench Score
68
Jun 16, 2026
71Instruction Success Rate
68
May 11, 2026
88.57Success Rate (SR)
65
Jun 19, 2026
87.5Genuine-Followup Rate
65
May 29, 2026
3,213Helpfulness Score
65
May 8, 2026
6.83Score
60
Feb 26, 2026
7.42Reasoning Score
60
Apr 6, 2026
19.84Accuracy
57
May 7, 2026
91.4Success Rate
55
Jul 7, 2026
81.13Score
55
Apr 10, 2026
96.3IFS
55
Feb 26, 2026
85.22Overall Score
55
Jun 16, 2026
98.9Accuracy (Greedy)
52
Jul 7, 2026
61.2MT-Bench Score
52
Feb 26, 2026
6.27Average Accepted Length
51
May 13, 2026
27.74Rouge-L
50
May 21, 2026
23.4R-L Score
50
Feb 26, 2026
88.92Avg. Score (IFEval)
49
Jun 26, 2026
80.9Win Rate
49
May 18, 2026
16.5ROUGE-L
48
Mar 5, 2026
83.1Score
46
May 29, 2026
54ROUGE-L
43
Mar 5, 2026
21.81ROUGE-L
42
Mar 5, 2026
29.69ROUGE-L
42
Mar 5, 2026