Grok
Benchmarks
TaskDatasetSOTA resultTrendResults
Grok 4
82NR Score
20
Grok Out-of-distribution (OOD) unseen domains (Legal, Email, Complaints) 4.1 (test)
99Legal Accuracy
16
92 Queries Grok 3 Evaluation 1.0 (test)
7.53Comprehensiveness
3
depth-8 grok true-MC
4.6Lambda Min (F-bar) Off Floor
2
grok grk_decompress_fuzzer
3,721Average Branch Coverage
2