Noometry

Model comparison

GPT-4 vs QwQ-32B

QwQ-32B is the stronger model overall, scoring 39.8 to 29.1 on the Noometry Index.

Last verified . 25 shared benchmarks.

GPT-4 OpenAI

29.1

Rank #316 Confirmed

QwQ-32B Alibaba (Qwen)

39.8

Rank #159 Confirmed

Summary

  • They share 25 benchmarks with published results for both. GPT-4 scores higher in 0 categories and QwQ-32B in 8 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where QwQ-32B leads 38.0 to 10.8.
  • The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.1% for GPT-4 and 59.2% for QwQ-32B.
  • QwQ-32B has downloadable open weights; the other is API-only.

Side by side

GPT-4QwQ-32B
ProviderOpenAIAlibaba (Qwen)
Noometry Index29.139.8
Released2023-03-142024-11-28
WeightsProprietaryOpen
Context window8K—
Max output8K—
Input $ / M tokens$30—
Output $ / M tokens$60—
Results tracked3836

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

Category by category

Coding QwQ-32B leads

GPT-4: 31.6 (#283), QwQ-32B: 35.4 (#226)

BenchmarkGPT-4QwQ-32B
BigCodeBench Instruct46%44.6%
LMArena Coding12541333
BigCodeBench Complete57.2%54.4%
Aider Polyglot—20.9%
WeirdML12.4%—
LiveBench Coding—72.2%
HumanEval+79.3%—

Agentic & Tool Use Not comparable

GPT-4: —, QwQ-32B: —

BenchmarkGPT-4QwQ-32B
METR Time Horizons36.1%—

Reasoning QwQ-32B leads

GPT-4: 17.8 (#289), QwQ-32B: 23.7 (#174)

BenchmarkGPT-4QwQ-32B
Chess Puzzles4%5%
LMArena Hard Prompts12411325
Epoch Capabilities Index125.89137.6
ForecastBench57.858.3
LiveBench Reasoning—83.5%
Mystery Game Puzzles12%—
DTBench62.7%—
LiveBench Data Analysis—65%
LMCA17.1%—
BIG-Bench Hard75.1%—
HellaSwag95.3%—
LiveBench—72%
WinoGrande87.5%—

Math QwQ-32B leads

GPT-4: 10.8 (#309), QwQ-32B: 38.0 (#143)

BenchmarkGPT-4QwQ-32B
OTIS Mock AIME 2024-20251.1%59.2%
LMArena Math12691359
LiveBench Math—77.8%
MATH Level 523%—
GSM8K92%—

Knowledge QwQ-32B leads

GPT-4: 18.4 (#282), QwQ-32B: 37.2 (#158)

BenchmarkGPT-4QwQ-32B
GPQA Diamond35.7%65.3%
LMArena Expert12111324
Confabulations—15.6%
MMLU86.4%—
TriviaQA84.8%—

Multilingual QwQ-32B leads

GPT-4: 40.6 (#215), QwQ-32B: 44.8 (#176)

BenchmarkGPT-4QwQ-32B
LMArena Non-English12461305
LMArena Chinese12421378
LMArena French12831336
LMArena German12511313
LMArena Japanese12091262
LMArena Korean11841279
LMArena Russian12511297
LMArena Spanish12611354

Instruction Following QwQ-32B leads

GPT-4: 65.3 (#222), QwQ-32B: 72.6 (#137)

BenchmarkGPT-4QwQ-32B
LMArena Instruction Following12411297
LiveBench Instruction Following—81.8%

Long Context QwQ-32B leads

GPT-4: 37.7 (#212), QwQ-32B: 49.0 (#11)

BenchmarkGPT-4QwQ-32B
LMArena Longer Query12441308
Fiction.LiveBench—83.3%

Writing & Preference QwQ-32B leads

GPT-4: 34.9 (#268), QwQ-32B: 50.6 (#180)

BenchmarkGPT-4QwQ-32B
LMArena Text12631329
LMArena Creative Writing12441288
EQ-Bench Creative Writing7521257
LMArena Multi-Turn12571314
Short-Story Creative Writing—80.2%
LiveBench Language—51.4%

Frequently asked questions

Is GPT-4 better than QwQ-32B?

QwQ-32B is the stronger model overall, scoring 39.8 to 29.1 on the Noometry Index.

Is GPT-4 or QwQ-32B better for coding?

QwQ-32B scores higher on coding benchmarks: 35.4 versus 31.6 in the Noometry coding category.

How many benchmarks do GPT-4 and QwQ-32B share?

25 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and QwQ-32B has 36.

Related comparisons

Go deeper