Noometry

Model comparison

GPT-5.4 nano vs Grok 4

Grok 4 is the stronger model overall, scoring 48.1 to 41.9 on the Noometry Index.

Last verified . 29 shared benchmarks.

GPT-5.4 nano OpenAI

41.9

Rank #125 Confirmed

Grok 4 xAI

48.1

Rank #56 Confirmed

Summary

  • They share 29 benchmarks with published results for both. GPT-5.4 nano scores higher in 1 category and Grok 4 in 8 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in long context, where Grok 4 leads 63.1 to 41.6.
  • The biggest single-benchmark swing is Kagi LLM Benchmark: 39.7% for GPT-5.4 nano and 73.6% for Grok 4.

Side by side

GPT-5.4 nanoGrok 4
ProviderOpenAIxAI
Noometry Index41.948.1
Released2026-03-172025-07-09
WeightsProprietaryProprietary
Context window400K—
Max output128K—
Input $ / M tokens$0.20—
Output $ / M tokens$1.25—
Results tracked4048

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

Category by category

Coding Grok 4 leads

GPT-5.4 nano: 43.6 (#84), Grok 4: 50.3 (#46)

BenchmarkGPT-5.4 nanoGrok 4
WeirdML49.2%45.7%
LMArena Coding14051408
Aider Polyglot—79.6%
SciCode46.9%—
ALE-Bench1,005—

Agentic & Tool Use Not comparable

GPT-5.4 nano: —, Grok 4: 32.3 (#68)

BenchmarkGPT-5.4 nanoGrok 4
Terminal-Bench—27.2%
Berkeley Function Calling Leaderboard—63%
GDPval—21.1%
Cybench—43%
DeepResearch Bench—47.3%
BALROG—43.6%
LMArena Search—1142
METR Time Horizons—66.6%

Reasoning Grok 4 leads

GPT-5.4 nano: 23.7 (#173), Grok 4: 36.7 (#65)

BenchmarkGPT-5.4 nanoGrok 4
ARC-AGI-25.7%16%
Kagi LLM Benchmark39.7%73.6%
ARC-AGI-151.5%66.7%
Chess Puzzles30%28%
LMArena Hard Prompts13811409
Epoch Capabilities Index145.81146.44
ForecastBench57.360.9
SimpleBench—60.5%
CritPt9.3%—
Mystery Game Puzzles9%—
DTBench80.3%—
LMCA36.9%—

Math Grok 4 leads

GPT-5.4 nano: 40.9 (#88), Grok 4: 48.4 (#64)

BenchmarkGPT-5.4 nanoGrok 4
OTIS Mock AIME 2024-202587.8%84%
LMArena Math14061422
FrontierMath (Feb 2025 set)25.9%19.7%
FrontierMath Tier 4 (v1)6.3%2.1%
FrontierMath (Tiers 1-3)44.9%—
FrontierMath Tier 412.2%—
ProofBench5%—
Omni-MATH—60.3%

Knowledge Grok 4 leads

GPT-5.4 nano: 41.9 (#103), Grok 4: 53.8 (#55)

BenchmarkGPT-5.4 nanoGrok 4
GPQA Diamond78.5%87%
LMArena Expert13961415
SimpleQA Verified11.7%—
MMLU-Pro—85.1%
Confabulations—12.4%
Vectara Hallucination Rate3.1%—
GPQA (HELM)—72.7%

Multimodal GPT-5.4 nano leads

GPT-5.4 nano: 36.7 (#78), Grok 4: 33.7 (#94)

BenchmarkGPT-5.4 nanoGrok 4
LMArena Vision11961210
GeoBench—45%

Multilingual Grok 4 leads

GPT-5.4 nano: 48.6 (#140), Grok 4: 51.8 (#103)

BenchmarkGPT-5.4 nanoGrok 4
LMArena Non-English13591403
LMArena Chinese13921427
LMArena French13961418
LMArena German13671429
LMArena Japanese13431394
LMArena Korean13201377
LMArena Russian13631410
LMArena Spanish13711420

Instruction Following Grok 4 leads

GPT-5.4 nano: 71.9 (#144), Grok 4: 79.2 (#5)

BenchmarkGPT-5.4 nanoGrok 4
LMArena Instruction Following13621387
IFEval—94.9%

Long Context Grok 4 leads

GPT-5.4 nano: 41.6 (#137), Grok 4: 63.1 (#4)

BenchmarkGPT-5.4 nanoGrok 4
LMArena Longer Query13661409
Fiction.LiveBench—94.4%

Writing & Preference Grok 4 leads

GPT-5.4 nano: 55.7 (#142), Grok 4: 58.5 (#116)

BenchmarkGPT-5.4 nanoGrok 4
LMArena Text13721411
LMArena Creative Writing13141397
LMArena Multi-Turn13821416
Short-Story Creative Writing—76.9%
WildBench—79.7%

Frequently asked questions

Is GPT-5.4 nano better than Grok 4?

Grok 4 is the stronger model overall, scoring 48.1 to 41.9 on the Noometry Index.

Is GPT-5.4 nano or Grok 4 better for coding?

Grok 4 scores higher on coding benchmarks: 50.3 versus 43.6 in the Noometry coding category.

How many benchmarks do GPT-5.4 nano and Grok 4 share?

29 benchmarks have published results for both models. GPT-5.4 nano has 40 scored results on Noometry and Grok 4 has 48.

Related comparisons

Go deeper