Instruction Following
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
8.13MT-Bench Score
10
Feb 26, 2026
4.25Helpfulness
10
Feb 26, 2026
4.08GPT-score
9
Jul 9, 2026
87.9Score
9
Jul 2, 2026
97.4Score
9
Jul 2, 2026
29.04Accuracy (Alpaca)
9
Jun 30, 2026
6.55MT-Bench Delta Score
9
May 29, 2026
45.2IFEval Delta
9
May 29, 2026
23.88IFBench Score
9
May 11, 2026
57.88Average Score
9
May 11, 2026
83.2Score
9
May 11, 2026
37.97ROUGE-L
9
May 11, 2026
52.26ROUGE-L
9
May 11, 2026
0.508BRT Score
9
Apr 24, 2026
46.9BRT
9
Apr 24, 2026
54.1BRT Score
9
Apr 24, 2026
119,345# Train Data
9
Apr 16, 2026
47.1Throughput (TPS)
9
May 19, 2026
86.1Accuracy
9
Feb 26, 2026
78.61AlpacaEval 2.0 LC Score
9
May 14, 2026
23.6ISR
9
Feb 26, 2026
36.4Instruction Success Rate
9
Feb 26, 2026
51.4ISR
9
Feb 26, 2026
66.6ISR
9
Feb 26, 2026
81.3ISR
9
Feb 26, 2026