Noometry

Model comparison

GPT-4 Turbo vs Muse Spark

Muse Spark is the stronger model overall, scoring 50.6 to 30.5 on the Noometry Index.

Last verified . 20 shared benchmarks.

GPT-4 Turbo OpenAI

30.5

Rank #292 Confirmed

Muse Spark Meta

50.6

Rank #46 Confirmed

Summary

  • They share 20 benchmarks with published results for both. GPT-4 Turbo scores higher in 0 categories and Muse Spark in 9 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in knowledge, where Muse Spark leads 65.7 to 24.3.
  • The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.7% for GPT-4 Turbo and 88.9% for Muse Spark.

Side by side

GPT-4 TurboMuse Spark
ProviderOpenAIMeta
Noometry Index30.550.6
Released2023-11-062026-04-08
WeightsProprietaryProprietary
Context window128K—
Max output4K—
Input $ / M tokens$10—
Output $ / M tokens$30—
Results tracked3627

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

Category by category

Coding Muse Spark leads

GPT-4 Turbo: 33.8 (#249), Muse Spark: 46.2 (#69)

BenchmarkGPT-4 TurboMuse Spark
LMArena Coding12681481
SciCode—51.5%
WeirdML18%—
BigCodeBench Instruct48.2%—
BigCodeBench Complete58.2%—
HumanEval+86.6%—
MBPP+73.3%—

Agentic & Tool Use Not comparable

GPT-4 Turbo: —, Muse Spark: —

BenchmarkGPT-4 TurboMuse Spark
METR Time Horizons36.7%—

Reasoning Muse Spark leads

GPT-4 Turbo: 15.3 (#317), Muse Spark: 35.9 (#67)

BenchmarkGPT-4 TurboMuse Spark
LMArena Hard Prompts12511474
Epoch Capabilities Index127.25152.04
SimpleBench25.1%—
CritPt—11.3%
Chess Puzzles6%—
DTBench61.6%—
LMCA9.8%—
ForecastBench59.4—

Math Muse Spark leads

GPT-4 Turbo: 9.0 (#322), Muse Spark: 47.8 (#66)

BenchmarkGPT-4 TurboMuse Spark
OTIS Mock AIME 2024-20256.7%88.9%
LMArena Math12721455
FrontierMath (Tiers 1-3)0.7%—
ProofBench—17%
MATH Level 546.7%—
FrontierMath (Feb 2025 set)—39%
FrontierMath Tier 4 (v1)—14.6%

Knowledge Muse Spark leads

GPT-4 Turbo: 24.3 (#268), Muse Spark: 65.7 (#13)

BenchmarkGPT-4 TurboMuse Spark
GPQA Diamond46.6%89.8%
LMArena Expert12231457
Humanity's Last Exam—40.6%
Confabulations28.4%—
MMLU81.3%—

Multimodal Muse Spark leads

GPT-4 Turbo: 30.6 (#110), Muse Spark: 43.4 (#24)

BenchmarkGPT-4 TurboMuse Spark
LMArena Vision10901306
LMArena Document—1444

Multilingual Muse Spark leads

GPT-4 Turbo: 40.5 (#216), Muse Spark: 56.1 (#24)

BenchmarkGPT-4 TurboMuse Spark
LMArena Non-English12451464
LMArena Chinese12421509
LMArena French12761497
LMArena German12591497
LMArena Korean11871459
LMArena Russian12591466
LMArena Spanish12601472
LMArena Japanese1194—

Instruction Following Muse Spark leads

GPT-4 Turbo: 65.8 (#216), Muse Spark: 75.9 (#51)

BenchmarkGPT-4 TurboMuse Spark
LMArena Instruction Following12491442

Long Context Muse Spark leads

GPT-4 Turbo: 38.0 (#206), Muse Spark: 44.4 (#69)

BenchmarkGPT-4 TurboMuse Spark
LMArena Longer Query12541451

Writing & Preference Muse Spark leads

GPT-4 Turbo: 47.7 (#206), Muse Spark: 66.0 (#39)

BenchmarkGPT-4 TurboMuse Spark
LMArena Text12721474
LMArena Creative Writing12691459
LMArena Multi-Turn12671477

Frequently asked questions

Is GPT-4 Turbo better than Muse Spark?

Muse Spark is the stronger model overall, scoring 50.6 to 30.5 on the Noometry Index.

Is GPT-4 Turbo or Muse Spark better for coding?

Muse Spark scores higher on coding benchmarks: 46.2 versus 33.8 in the Noometry coding category.

How many benchmarks do GPT-4 Turbo and Muse Spark share?

20 benchmarks have published results for both models. GPT-4 Turbo has 36 scored results on Noometry and Muse Spark has 27.

Related comparisons

Go deeper