Noometry

Model comparison

GPT-4.5 vs Grok 4.5

Grok 4.5 is the stronger model overall, scoring 55.0 to 37.2 on the Noometry Index.

Last verified . 25 shared benchmarks.

GPT-4.5 OpenAI

37.2

Rank #208 Confirmed

Grok 4.5 xAI

55.0

Rank #25 Confirmed

Summary

  • They share 25 benchmarks with published results for both. GPT-4.5 scores higher in 1 category and Grok 4.5 in 9 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where Grok 4.5 leads 56.1 to 13.9.
  • The biggest single-benchmark swing is ARC-AGI-1: 10.3% for GPT-4.5 and 87.2% for Grok 4.5.

Side by side

GPT-4.5Grok 4.5
ProviderOpenAIxAI
Noometry Index37.255.0
Released2025-02-272026-07-08
WeightsProprietaryProprietary
Context window—500K
Max output—500K
Input $ / M tokens—$2
Output $ / M tokens—$6
Results tracked4252

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

Category by category

Coding Grok 4.5 leads

GPT-4.5: 42.2 (#109), Grok 4.5: 52.2 (#35)

BenchmarkGPT-4.5Grok 4.5
WeirdML39.4%46.4%
LMArena Coding13961474
DeepSWE—53.8%
FrontierCode—42.4%
Aider Polyglot44.9%—
LMArena WebDev—1553
SciCode—54.1%
LiveBench Coding75.2%—
ALE-Bench—1,309

Agentic & Tool Use Grok 4.5 leads

GPT-4.5: 27.9 (#97), Grok 4.5: 44.4 (#17)

BenchmarkGPT-4.5Grok 4.5
APEX-Agents—56.2%
τ²-bench Banking—47.9%
Cybench17.5%—
PostTrainBench—23.4%
GBAEval—65.4%
GDP.pdf—14%
LMArena Search—1213
Vending-Bench 2—3,887

Reasoning Grok 4.5 leads

GPT-4.5: 13.9 (#330), Grok 4.5: 56.1 (#25)

BenchmarkGPT-4.5Grok 4.5
ARC-AGI-20.8%52.6%
SimpleBench34.5%70%
ARC-AGI-110.3%87.2%
LMArena Hard Prompts14031462
Epoch Capabilities Index136.74153.92
Kagi LLM Benchmark—83.5%
NYT Connections (extended)—79.9%
CritPt—15.4%
Chess Puzzles—36%
EnigmaEval3.2%—
LiveBench Reasoning71.1%—
DTBench—96.5%
LiveBench Data Analysis64.3%—
LMCA—45.2%
Surface Evolver Bench—74.4%
ForecastBench61.7—
LiveBench69%—

Math Grok 4.5 leads

GPT-4.5: 32.6 (#211), Grok 4.5: 60.9 (#35)

BenchmarkGPT-4.5Grok 4.5
OTIS Mock AIME 2024-202537.8%97.8%
LMArena Math14121459
FrontierMath (Tiers 1-3)—57.2%
FrontierMath Tier 4—24.4%
ProofBench—31%
LiveBench Math69.3%—
MATH Level 578.6%—

Knowledge Grok 4.5 leads

GPT-4.5: 32.5 (#211), Grok 4.5: 62.3 (#24)

BenchmarkGPT-4.5Grok 4.5
GPQA Diamond68.7%93.4%
LMArena Expert13941466
Humanity's Last Exam5.4%—
SimpleQA Verified—48.3%
Confabulations13.6%—

Multimodal Too close to call

GPT-4.5: 37.6 (#71), Grok 4.5: 37.6 (#72)

BenchmarkGPT-4.5Grok 4.5
LMArena Vision11951288
VPCT45%—
Blueprint-Bench 2—27.3%
Furniture Assembly—22.5%
LMArena Document—1452

Multilingual Grok 4.5 leads

GPT-4.5: 52.5 (#83), Grok 4.5: 54.4 (#42)

BenchmarkGPT-4.5Grok 4.5
LMArena Non-English14131440
LMArena Chinese14211496
LMArena French14181456
LMArena German14571446
LMArena Japanese14161428
LMArena Korean13921404
LMArena Russian14191448
LMArena Spanish—1450

Instruction Following Grok 4.5 leads

GPT-4.5: 72.6 (#134), Grok 4.5: 76.0 (#48)

BenchmarkGPT-4.5Grok 4.5
LMArena Instruction Following14041446
LiveBench Instruction Following72.3%—

Long Context Grok 4.5 leads

GPT-4.5: 40.4 (#155), Grok 4.5: 44.8 (#56)

BenchmarkGPT-4.5Grok 4.5
LMArena Longer Query14061463
Fiction.LiveBench63.9%—

Writing & Preference Grok 4.5 leads

GPT-4.5: 56.9 (#134), Grok 4.5: 65.8 (#42)

BenchmarkGPT-4.5Grok 4.5
LMArena Text14171448
LMArena Creative Writing13941442
EQ-Bench Creative Writing12581579
LMArena Multi-Turn14441456
Short-Story Creative Writing75.6%—
LiveBench Language61.5%—

Frequently asked questions

Is GPT-4.5 better than Grok 4.5?

Grok 4.5 is the stronger model overall, scoring 55.0 to 37.2 on the Noometry Index.

Is GPT-4.5 or Grok 4.5 better for coding?

Grok 4.5 scores higher on coding benchmarks: 52.2 versus 42.2 in the Noometry coding category.

How many benchmarks do GPT-4.5 and Grok 4.5 share?

25 benchmarks have published results for both models. GPT-4.5 has 42 scored results on Noometry and Grok 4.5 has 52.

Related comparisons

Go deeper