xAIclosed weightsReasoning
Grok 4
Jul 9, 2025
$3 / $15
$6
256K
Benchmark scores
| Benchmark | Score | Δ vs prior | Rank (current) | Configuration | Source |
|---|---|---|---|---|---|
GPQA Diamond | 87.5% | #9 of 11 | lab-reported | x.ai ↗ | |
Humanity's Last Exam | 25.4% | #15 of 15 | no tools · lab-reported | x.ai ↗ | |
ARC-AGI-2 | 15.9% | #10 of 10 | lab-reported | x.ai ↗ | |
τ²-bench (Telecom) | 74.9% | #7 of 10 | leaderboard | benchlm.ai ↗ | |
Artificial Analysis Intelligence Index | 33.3 | #7 of 11 | leaderboard | benchlm.ai ↗ |