Model comparison

GPT-4 vs Qwen2.5 72B Instruct

Qwen2.5 72B Instruct is the stronger model overall, scoring 31.9 to 29.1 on the Noometry Index.

Last verified October 10, 2026. 33 shared benchmarks.

GPT-4 OpenAI

29.1

Rank #316 Confirmed

Qwen2.5 72B Instruct

Alibaba (Qwen)

31.9

Rank #267 Confirmed

Summary

They share 33 benchmarks with published results for both. GPT-4 scores higher in 0 categories and Qwen2.5 72B Instruct in 8 categories; 6 gaps are clear of the uncertainty.
The widest gap is in writing & preference, where Qwen2.5 72B Instruct leads 46.7 to 34.9.
The biggest single-benchmark swing is MATH Level 5: 23% for GPT-4 and 63.2% for Qwen2.5 72B Instruct.
Qwen2.5 72B Instruct is cheaper at $1.40 / $5.60 per million input/output tokens, against $30 / $60 for GPT-4.
Qwen2.5 72B Instruct accepts more context: 131K tokens versus 8K.
Qwen2.5 72B Instruct has downloadable open weights; the other is API-only.

Side by side

GPT-4 and Qwen2.5 72B Instruct specifications
	GPT-4	Qwen2.5 72B Instruct
Provider	OpenAI	Alibaba (Qwen)
Noometry Index	29.1	31.9
Released	2023-03-14	2024-09
Weights	Proprietary	Open
Context window	8K	131K
Max output	8K	8K
Input $ / M tokens	$30	$1.40
Output $ / M tokens	$60	$5.60
Results tracked	38	43

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

Category by category

Coding Qwen2.5 72B Instruct leads

GPT-4: 31.6 (#283), Qwen2.5 72B Instruct: 33.2 (#260)

Coding benchmarks
Benchmark	GPT-4	Qwen2.5 72B Instruct
WeirdML	12.4%	16%
BigCodeBench Instruct	46%	45.8%
LMArena Coding	1254	1292
BigCodeBench Complete	57.2%	55.9%
HumanEval+	79.3%	—

Agentic & Tool Use Not comparable

GPT-4: —, Qwen2.5 72B Instruct: 22.1 (#133)

Agentic & Tool Use benchmarks
Benchmark	GPT-4	Qwen2.5 72B Instruct
METR Time Horizons	36.1%	35.8%
TheAgentCompany	—	5.7%
BALROG	—	16.2%

Reasoning Qwen2.5 72B Instruct leads

GPT-4: 17.8 (#289), Qwen2.5 72B Instruct: 22.3 (#199)

Reasoning benchmarks
Benchmark	GPT-4	Qwen2.5 72B Instruct
LMArena Hard Prompts	1241	1271
DTBench	62.7%	62.9%
LMCA	17.1%	13.4%
BIG-Bench Hard	75.1%	79.8%
Epoch Capabilities Index	125.89	129
ForecastBench	57.8	57.5
HellaSwag	95.3%	84.8%
WinoGrande	87.5%	82.3%
Chess Puzzles	4%	—
Mystery Game Puzzles	12%	—
PIQA	—	82.6%

Math Qwen2.5 72B Instruct leads

GPT-4: 10.8 (#309), Qwen2.5 72B Instruct: 19.3 (#287)

Math benchmarks
Benchmark	GPT-4	Qwen2.5 72B Instruct
OTIS Mock AIME 2024-2025	1.1%	8.1%
LMArena Math	1269	1283
MATH Level 5	23%	63.2%
Omni-MATH	—	33%
GSM8K	92%	—

Knowledge Qwen2.5 72B Instruct leads

GPT-4: 18.4 (#282), Qwen2.5 72B Instruct: 27.0 (#253)

Knowledge benchmarks
Benchmark	GPT-4	Qwen2.5 72B Instruct
GPQA Diamond	35.7%	49.1%
LMArena Expert	1211	1245
MMLU	86.4%	85.3%
TriviaQA	84.8%	71.9%
MMLU-Pro	—	63.1%
Confabulations	—	19.1%
GPQA (HELM)	—	42.6%
ARC (AI2) Challenge	—	94.5%

Multilingual Too close to call

GPT-4: 40.6 (#215), Qwen2.5 72B Instruct: 41.0 (#213)

Multilingual benchmarks
Benchmark	GPT-4	Qwen2.5 72B Instruct
LMArena Non-English	1246	1252
LMArena Chinese	1242	1272
LMArena French	1283	1280
LMArena German	1251	1234
LMArena Japanese	1209	1180
LMArena Korean	1184	1188
LMArena Russian	1251	1264
LMArena Spanish	1261	1256

Instruction Following Too close to call

GPT-4: 65.3 (#222), Qwen2.5 72B Instruct: 65.5 (#221)

Instruction Following benchmarks
Benchmark	GPT-4	Qwen2.5 72B Instruct
LMArena Instruction Following	1241	1254
IFEval	—	80.6%

Long Context Qwen2.5 72B Instruct leads

GPT-4: 37.7 (#212), Qwen2.5 72B Instruct: 38.9 (#188)

Long Context benchmarks
Benchmark	GPT-4	Qwen2.5 72B Instruct
LMArena Longer Query	1244	1282

Writing & Preference Qwen2.5 72B Instruct leads

GPT-4: 34.9 (#268), Qwen2.5 72B Instruct: 46.7 (#215)

Writing & Preference benchmarks
Benchmark	GPT-4	Qwen2.5 72B Instruct
LMArena Text	1263	1269
LMArena Creative Writing	1244	1221
LMArena Multi-Turn	1257	1272
EQ-Bench Creative Writing	752	—
WildBench	—	80.2%

Frequently asked questions

Is GPT-4 better than Qwen2.5 72B Instruct?

Qwen2.5 72B Instruct is the stronger model overall, scoring 31.9 to 29.1 on the Noometry Index.

Which is cheaper, GPT-4 or Qwen2.5 72B Instruct?

Qwen2.5 72B Instruct is cheaper. It lists at $1.40 per million input tokens and $5.60 per million output tokens; GPT-4 lists at $30 and $60.

Is GPT-4 or Qwen2.5 72B Instruct better for coding?

Qwen2.5 72B Instruct scores higher on coding benchmarks: 33.2 versus 31.6 in the Noometry coding category.

Which has the bigger context window?

Qwen2.5 72B Instruct does, with 131K tokens against 8K.

How many benchmarks do GPT-4 and Qwen2.5 72B Instruct share?

33 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Qwen2.5 72B Instruct has 43.