Grok 2 mini: benchmark results

Provider: xAI. Access: API.

Unified ELO 1574 ± 42, rank #733 of 2656 rated models, from 9 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
LLM Stats (MathVista)68.1Score (%)53.9
LLM Stats (DocVQA)93.2Score (%)53.7
BenchTable - Reasoning35.3Weighted Score (%)43
BenchTable - Tech48.1Weighted Score (%)41.6
BenchTable42.6Total Score (%)40.7
BenchTable - STEM35.7Weighted Score (%)31.7
BenchTable - Utility38.4Weighted Score (%)28.3
LLM Stats Score8.58LLM Stats Score (conservative rating)25.3
ZeroEval GPQA Diamond51GPQA Diamond Score24.1

Interactive version: theaggregate.ai/model?slug=grok-2-mini · How It Works · Data refreshed daily, snapshot 2026-09-19.