Grok 4.5 (High) — benchmark results

Provider: xAI. Access: API.

Unified ELO 1984 ± 30, rank #22 of 1719 rated models, from 35 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Tau3 Banking32.58Success Rate (%)100
AA GPQA Diamond93.13Accuracy (%)99.4
Artificial Analysis Intelligence Index53.83Intelligence Index99.4
AA Omniscience26.38Score99.3
AA Omniscience - Health48.5Accuracy (%)99.3
AA Omniscience - Science, Engineering & Mathematics51.8Accuracy (%)99
AA Omniscience - Software Engineering (SWE) - R70Accuracy (%)98.6
AA Humanity's Last Exam40.27Accuracy (%)98.4
AA SciCode54.05Accuracy (%)98.4
AA-Omniscience Accuracy52.05Accuracy (%)98.1
AA Omniscience - Software Engineering (SWE) - Go72Accuracy (%)97.7
AA Omniscience - Law50.1Accuracy (%)97.6

Interactive version: aibenchmarks.dev/model?slug=grok-4-5-high · How the rankings work · Data refreshed daily, snapshot 2026-07-09.