Noometry

Model comparison

Grok 3 vs MiniMax-M2.7

Grok 3 is the stronger model overall, scoring 39.9 to 37.7 on the Noometry Index.

Last verified . 20 shared benchmarks.

Grok 3 xAI

39.9

Rank #157 Confirmed

MiniMax-M2.7 MiniMax

37.7

Rank #196 Confirmed

Summary

  • They share 20 benchmarks with published results for both. Grok 3 scores higher in 6 categories and MiniMax-M2.7 in 3 categories; 7 gaps are clear of the uncertainty.
  • The widest gap is in math, where Grok 3 leads 38.0 to 25.9.
  • The biggest single-benchmark swing is Vectara Hallucination Rate: 5.8% for Grok 3 and 12.9% for MiniMax-M2.7.
  • MiniMax-M2.7 has downloadable open weights; the other is API-only.

Side by side

Grok 3MiniMax-M2.7
ProviderxAIMiniMax
Noometry Index39.937.7
Released2025-04-092026-03-18
WeightsProprietaryOpen
Context window—205K
Max output—131K
Input $ / M tokens—$0.30
Output $ / M tokens—$1.20
Results tracked4030

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Too close to call

Grok 3: 41.9 (#115), MiniMax-M2.7: 41.8 (#120)

BenchmarkGrok 3MiniMax-M2.7
WeirdML37.2%37%
LMArena Coding14321454
Aider Polyglot53.3%—
LMArena WebDev—1398
SciCode—47%
ALE-Bench—599.25

Agentic & Tool Use Grok 3 leads

Grok 3: 30.5 (#76), MiniMax-M2.7: 25.1 (#111)

BenchmarkGrok 3MiniMax-M2.7
Terminal-Bench—45.1%
BALROG29.5%—
ExploitBench—13.3%
GBAEval—0%

Reasoning MiniMax-M2.7 leads

Grok 3: 13.7 (#333), MiniMax-M2.7: 19.7 (#253)

BenchmarkGrok 3MiniMax-M2.7
LMArena Hard Prompts14341422
Epoch Capabilities Index138.33145.85
ARC-AGI-20%—
SimpleBench36.1%—
Kagi LLM Benchmark61.3%—
NYT Connections (extended)—24.7%
ARC-AGI-15.5%—
CritPt—0.6%
Thematic Generalization—39.3%

Math Grok 3 leads

Grok 3: 38.0 (#145), MiniMax-M2.7: 25.9 (#263)

BenchmarkGrok 3MiniMax-M2.7
LMArena Math13911420
OTIS Mock AIME 2024-202555.6%—
ProofBench—3%
Omni-MATH46.4%—
MATH Level 588.7%—
FrontierMath (Feb 2025 set)3.8%—
FrontierMath Tier 4 (v1)0%—

Knowledge Grok 3 leads

Grok 3: 46.2 (#82), MiniMax-M2.7: 37.7 (#152)

BenchmarkGrok 3MiniMax-M2.7
Vectara Hallucination Rate5.8%12.9%
LMArena Expert14211444
GPQA Diamond75.8%—
MMLU-Pro78.8%—
Confabulations14.2%—
GPQA (HELM)65%—

Multilingual Grok 3 leads

Grok 3: 52.3 (#87), MiniMax-M2.7: 50.3 (#123)

BenchmarkGrok 3MiniMax-M2.7
LMArena Non-English14101382
LMArena Chinese14481441
LMArena French14601421
LMArena German14311398
LMArena Japanese13871262
LMArena Korean13731313
LMArena Russian14161383
LMArena Spanish14171403

Instruction Following Too close to call

Grok 3: 75.0 (#73), MiniMax-M2.7: 74.1 (#103)

BenchmarkGrok 3MiniMax-M2.7
LMArena Instruction Following14091405
IFEval88.4%—

Long Context MiniMax-M2.7 leads

Grok 3: 38.7 (#192), MiniMax-M2.7: 43.3 (#99)

BenchmarkGrok 3MiniMax-M2.7
LMArena Longer Query14391419
Fiction.LiveBench58.3%—

Writing & Preference MiniMax-M2.7 leads

Grok 3: 55.8 (#141), MiniMax-M2.7: 58.9 (#112)

BenchmarkGrok 3MiniMax-M2.7
LMArena Text14261405
LMArena Creative Writing14141354
LMArena Multi-Turn14251412
Short-Story Creative Writing76.4%—
EQ-Bench Creative Writing1186—
WildBench84.9%—

Frequently asked questions

Is Grok 3 better than MiniMax-M2.7?

Grok 3 is the stronger model overall, scoring 39.9 to 37.7 on the Noometry Index.

Is Grok 3 or MiniMax-M2.7 better for coding?

They score almost the same on coding (41.9 vs 41.8); test both on your own repository before choosing.

How many benchmarks do Grok 3 and MiniMax-M2.7 share?

20 benchmarks have published results for both models. Grok 3 has 40 scored results on Noometry and MiniMax-M2.7 has 30.

Related comparisons

Go deeper