Grok 4 Fast (Reasoning) — benchmark results

Grok 4 Fast evaluated with reasoning enabled. Provider: xAI. Released 2025-09-19. Access: API.

Unified ELO 1709 ± 12, rank #259 of 1776 rated models, from 127 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
NYT Connections Older Models92.2Score (%)98.1
LiveOIBench69.8Avg Human Percentile96.5
SpeechMap Compliance94.3% Requests Completed96.4
AA LiveCodeBench83.17Pass@1 (%)94.4
UGI Leaderboard51.29UGI Score92.2
Vals AI TaxEval v275.7Accuracy (%)92.2
Diplomacy: Overall Performance57.75Score90.9
AA MMLU-Pro85.04Accuracy (%)90.8
UGI - Writing53.37Writing Score90.5
Vals AI CorpFin v266.9Accuracy (%)90.2
AA AIME 202589.67Accuracy (%)89.8
UGI - Natural Intelligence46.53NatInt Score89.4

Interactive version: theaggregate.ai/model?slug=grok-4-fast-reasoning · How the rankings work · Data refreshed daily, snapshot 2026-07-22.