Models/Grok 3

MDL

Grok 3

xAI

Verified against Artificial Analysis · Jul 13, 2026

Capability

percentile index

Price /M

$4 / $20

in / out

Context

—

Avg score

49.1

11 benchmarks

AA Intelligence

18.4

index

Weights

Closed

Released

Feb 19, 2025

About

Grok 3 is a proprietary model in the Grok family from xAI. Benchmarked on 11 evals, averaging 49.1.

Benchmark scores · 11

avg 49.1 — higher is better

MATHartificial-analysis · 2026-07

MMLU-Proartificial-analysis · 2026-07

79.9

GPQA Diamondartificial-analysis · 2026-07

69.3

AIMEartificial-analysis · 2026-07

AA Long-Context Reasoningartificial-analysis · 2026-07

54.7

τ²-benchartificial-analysis · 2026-07

48.8

IFBenchartificial-analysis · 2026-07

46.9

LiveCodeBenchartificial-analysis · 2026-07

42.5

SciCodeartificial-analysis · 2026-07

36.8

Terminal-Bench Hardartificial-analysis · 2026-07

11.4

Humanity's Last Examartificial-analysis · 2026-07

5.1

Details

PublisherxAI

FamilyGrok

Parameters—

License—

WeightsClosed

ReleasedFeb 19, 2025

Performance & price

Price in / out$4 / $20 per 1M

Blended price$8 / 1M

Math index58.0

Median across API providers, via Artificial Analysis. Blended = 3:1 input:output. For reasoning models, latency includes thinking time. Methodology →

History

full ledger →

Capability over time

Append-only ledger — every observed change to this model's numbers.

API providers

Artificial Analysis$20/M