benchgap
xAI · model

Grok 3 [Beta] benchmark scores

As of 2026-11-22, Grok 3 [Beta] (xAI) has measured scores on 5 benchmarks and estimated scores on 22 more.

BenchmarkScoreSource
AA-GPQA Diamond73.4%estimated ± 6.2 pp, medium confidence
AA-HLE12.0%estimated ± 5.7 pp, medium confidence
Artificial Analysis Intelligence Index11.1%estimated ± 7.4 pp, medium confidence
CritPt0.4%estimated ± 4.0 pp, high confidence
GPQA Diamond75.6%estimated ± 1.2 pp, high confidence
HLE12.6%estimated ± 6.5 pp, medium confidence
AA-Omniscience Accuracy31.3%estimated ± 2.7 pp, low confidence
SWE-bench Verified56.3%estimated ± 9.3 pp, low confidence
MMLU-Pro79.9%measured
AIME 202452.2%measured
BBH95.9%estimated ± 2.8 pp, medium confidence
GPQA75.4%measured
MMLU88.6%estimated ± 2.1 pp, high confidence
Vals GPQA Diamond73.3%estimated ± 4.5 pp, high confidence
Vals MMLU-Pro81.2%estimated ± 2.3 pp, high confidence
MATH-50093.5%estimated ± 1.3 pp, high confidence
MMMLU85.2%estimated ± 1.7 pp, high confidence
AIME 202537.1%estimated ± 3.5 pp, high confidence
AA MMLU-Pro85.7%estimated ± 1.2 pp, low confidence
ARC-AGI-136.9%estimated ± 6.7 pp, low confidence
ARC-AGI-20.0%estimated ± 10.9 pp, low confidence
HLE w/o tools12.0%estimated ± 5.3 pp, medium confidence
C-Eval89.8%estimated ± 1.2 pp, high confidence
MMLU-Redux90.3%estimated ± 0.6 pp, high confidence
SuperGPQA55.4%estimated ± 8.5 pp, low confidence
SimpleQA43.6%measured
LiveCodeBench57.0%measured