Noometry

Model comparison

GPT-5.4 nano vs Grok-3 mini

GPT-5.4 nano and Grok-3 mini score almost the same on the Noometry Index (41.9 vs 41.2), so choose on price, context window or the category you care about most.

Last verified . 25 shared benchmarks.

GPT-5.4 nano OpenAI

41.9

Rank #125 Confirmed

Grok-3 mini xAI

41.2

Rank #141 Confirmed

Summary

  • They share 25 benchmarks with published results for both. GPT-5.4 nano scores higher in 5 categories and Grok-3 mini in 3 categories; 6 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where GPT-5.4 nano leads 23.7 to 13.6.
  • The biggest single-benchmark swing is ARC-AGI-1: 51.5% for GPT-5.4 nano and 16.5% for Grok-3 mini.

Side by side

GPT-5.4 nanoGrok-3 mini
ProviderOpenAIxAI
Noometry Index41.941.2
Released2026-03-172025-04-09
WeightsProprietaryProprietary
Context window400K—
Max output128K—
Input $ / M tokens$0.20—
Output $ / M tokens$1.25—
Results tracked4035

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

Category by category

Coding GPT-5.4 nano leads

GPT-5.4 nano: 43.6 (#84), Grok-3 mini: 40.8 (#131)

BenchmarkGPT-5.4 nanoGrok-3 mini
WeirdML49.2%42.6%
LMArena Coding14051379
Aider Polyglot—49.3%
SciCode46.9%—
ALE-Bench1,005—

Reasoning GPT-5.4 nano leads

GPT-5.4 nano: 23.7 (#173), Grok-3 mini: 13.6 (#334)

BenchmarkGPT-5.4 nanoGrok-3 mini
ARC-AGI-25.7%0.4%
Kagi LLM Benchmark39.7%61.3%
ARC-AGI-151.5%16.5%
LMArena Hard Prompts13811375
Epoch Capabilities Index145.81140.35
CritPt9.3%—
Chess Puzzles30%—
Mystery Game Puzzles9%—
DTBench80.3%—
LMCA36.9%—
ForecastBench57.3—

Math Grok-3 mini leads

GPT-5.4 nano: 40.9 (#88), Grok-3 mini: 42.1 (#85)

BenchmarkGPT-5.4 nanoGrok-3 mini
OTIS Mock AIME 2024-202587.8%77.8%
LMArena Math14061386
FrontierMath (Feb 2025 set)25.9%5.9%
FrontierMath (Tiers 1-3)44.9%—
FrontierMath Tier 412.2%—
ProofBench5%—
Omni-MATH—31.8%
MATH Level 5—90.9%
FrontierMath Tier 4 (v1)6.3%—

Knowledge Grok-3 mini leads

GPT-5.4 nano: 41.9 (#103), Grok-3 mini: 46.4 (#81)

BenchmarkGPT-5.4 nanoGrok-3 mini
GPQA Diamond78.5%76.3%
LMArena Expert13961395
SimpleQA Verified11.7%—
MMLU-Pro—79.9%
Confabulations—10.8%
Vectara Hallucination Rate3.1%—
GPQA (HELM)—67.5%

Multimodal Not comparable

GPT-5.4 nano: 36.7 (#78), Grok-3 mini: —

BenchmarkGPT-5.4 nanoGrok-3 mini
LMArena Vision1196—

Multilingual Too close to call

GPT-5.4 nano: 48.6 (#140), Grok-3 mini: 48.1 (#145)

BenchmarkGPT-5.4 nanoGrok-3 mini
LMArena Non-English13591352
LMArena Chinese13921387
LMArena French13961357
LMArena German13671349
LMArena Japanese13431342
LMArena Korean13201335
LMArena Russian13631353
LMArena Spanish13711381

Instruction Following Grok-3 mini leads

GPT-5.4 nano: 71.9 (#144), Grok-3 mini: 78.5 (#9)

BenchmarkGPT-5.4 nanoGrok-3 mini
LMArena Instruction Following13621357
IFEval—95.1%

Long Context Too close to call

GPT-5.4 nano: 41.6 (#137), Grok-3 mini: 41.0 (#147)

BenchmarkGPT-5.4 nanoGrok-3 mini
LMArena Longer Query13661372
Fiction.LiveBench—66.7%

Writing & Preference GPT-5.4 nano leads

GPT-5.4 nano: 55.7 (#142), Grok-3 mini: 52.5 (#169)

BenchmarkGPT-5.4 nanoGrok-3 mini
LMArena Text13721370
LMArena Creative Writing13141342
LMArena Multi-Turn13821355
Short-Story Creative Writing—73.5%
WildBench—65.1%

Frequently asked questions

Is GPT-5.4 nano better than Grok-3 mini?

GPT-5.4 nano and Grok-3 mini score almost the same on the Noometry Index (41.9 vs 41.2), so choose on price, context window or the category you care about most.

Is GPT-5.4 nano or Grok-3 mini better for coding?

GPT-5.4 nano scores higher on coding benchmarks: 43.6 versus 40.8 in the Noometry coding category.

How many benchmarks do GPT-5.4 nano and Grok-3 mini share?

25 benchmarks have published results for both models. GPT-5.4 nano has 40 scored results on Noometry and Grok-3 mini has 35.

Related comparisons

Go deeper