Instruction Following
Benchmarks
Dataset NameSOTA methodMetricTrendResultsLast Updated
2,688LLM Throughput (tokens/s)
8
May 12, 2026
83.9MT-Bench Score
8
Apr 2, 2026
9.6Score
8
Feb 26, 2026
78.22Accuracy
8
Feb 26, 2026
39.2IF Score
8
Feb 26, 2026
55.3Spatial Success Rate
7
Jun 23, 2026
89.53Accuracy
7
Jun 23, 2026
88.97Accuracy
7
Jun 23, 2026
40.1Lexical Dominance
7
Jun 2, 2026
29.16Task Success Rate
7
May 26, 2026
57.6AA
7
May 12, 2026
12.91ROUGE Score
7
Apr 27, 2026
65Exact Match (EM)
7
Mar 20, 2026
0.361Average Episodic Reward
7
Feb 26, 2026
0.486Average Episodic Reward
7
Feb 26, 2026
0.575Average Episodic Reward
7
Feb 26, 2026
82.97Task Performance
7
Feb 26, 2026
90.4Accuracy
7
May 14, 2026
83.6Ko-IFEval
7
Feb 26, 2026
67.7Average Win Rate
7
Feb 26, 2026
58.3Win Rate (bn)
7
Feb 26, 2026
62.8Win Rate (bn)
7
Feb 26, 2026
81.7Win Rate (bn)
7
Feb 26, 2026
0.85Win Rate (bn)
7
Feb 26, 2026
99.51Accuracy
7
Feb 26, 2026