Model comparison

DeepSeek-V3.2-Exp vs GPT-4o

DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 28.6 on the Noometry Index.

Last verified October 10, 2026. 33 shared benchmarks.

DeepSeek-V3.2-Exp

DeepSeek

44.3

Rank #78 Confirmed

GPT-4o OpenAI

28.6

Rank #324 Confirmed

Summary

They share 33 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 9 categories and GPT-4o in 0 categories; 9 gaps are clear of the uncertainty.
The widest gap is in math, where DeepSeek-V3.2-Exp leads 41.7 to 10.6.
The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 87.8% for DeepSeek-V3.2-Exp and 6.4% for GPT-4o.
DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $2.50 / $10 for GPT-4o.
DeepSeek-V3.2-Exp accepts more context: 164K tokens versus 128K.
DeepSeek-V3.2-Exp has downloadable open weights; the other is API-only.

Side by side

DeepSeek-V3.2-Exp and GPT-4o specifications
	DeepSeek-V3.2-Exp	GPT-4o
Provider	DeepSeek	OpenAI
Noometry Index	44.3	28.6
Released	2025-09-29	2024-05-13
Weights	Open	Proprietary
Context window	164K	128K
Max output	66K	16K
Input $ / M tokens	$0.26	$2.50
Output $ / M tokens	$0.38	$10
Results tracked	49	72

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

Category by category

Coding DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 46.5 (#65), GPT-4o: 24.8 (#328)

Coding benchmarks
Benchmark	DeepSeek-V3.2-Exp	GPT-4o
SWE-bench Verified (bash only)	70%	21.6%
Aider Polyglot	74.2%	45.3%
WeirdML	39.5%	25.1%
LMArena Coding	1454	1297
SWE-bench Verified	—	31%
LMArena WebDev	1362	—
SWE-bench Multilingual	59%	—
SciCode	38.9%	—
GSO	—	0%
BigCodeBench Instruct	—	51.1%
LiveBench Coding	—	51.4%
BigCodeBench Complete	—	61.1%
CadEval	—	26%
HumanEval+	—	87.2%
MBPP+	—	72.2%

Agentic & Tool Use DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 32.7 (#59), GPT-4o: 21.0 (#141)

Agentic & Tool Use benchmarks
Benchmark	DeepSeek-V3.2-Exp	GPT-4o
TheAgentCompany	42.9%	8.6%
Terminal-Bench	39.6%	—
APEX-Agents	21.3%	—
Berkeley Function Calling Leaderboard	56.7%	—
GDPval	—	9.9%
Cybench	—	12.5%
BALROG	—	32.3%
LMArena Search	—	1006
METR Time Horizons	—	40.8%
Vending-Bench 2	1,034	—

Reasoning DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 22.1 (#208), GPT-4o: 9.4 (#343)

Reasoning benchmarks
Benchmark	DeepSeek-V3.2-Exp	GPT-4o
ARC-AGI-2	4%	0%
ARC-AGI-1	57%	4.5%
CritPt	2.9%	0%
Chess Puzzles	14%	13%
LMArena Hard Prompts	1434	1281
DTBench	87.7%	64.5%
LMCA	29.1%	16.6%
Epoch Capabilities Index	146.27	128.97
SimpleBench	—	17.8%
Kagi LLM Benchmark	52.2%	—
NYT Connections (extended)	36.7%	—
EnigmaEval	—	0.8%
Thematic Generalization	65%	—
LiveBench Reasoning	—	55.8%
LiveBench Data Analysis	—	60.9%
ForecastBench	—	57.7
LiveBench	—	55.3%

Math DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 41.7 (#87), GPT-4o: 10.6 (#312)

Math benchmarks
Benchmark	DeepSeek-V3.2-Exp	GPT-4o
OTIS Mock AIME 2024-2025	87.8%	6.4%
LMArena Math	1435	1285
FrontierMath (Feb 2025 set)	22.1%	0.3%
FrontierMath (Tiers 1-3)	—	0.4%
MathArena Final-Answer Competitions	57.7%	—
ProofBench	8%	—
Omni-MATH	—	29.3%
LiveBench Math	—	49.5%
MATH Level 5	—	53.3%
FrontierMath Tier 4 (v1)	2.1%	—

Knowledge DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 51.7 (#66), GPT-4o: 28.8 (#242)

Knowledge benchmarks
Benchmark	DeepSeek-V3.2-Exp	GPT-4o
GPQA Diamond	83.4%	49.2%
Vectara Hallucination Rate	5.3%	9.6%
LMArena Expert	1436	1250
Humanity's Last Exam	—	2.7%
SimpleQA Verified	—	26%
MMLU-Pro	—	71.3%
Confabulations	—	15.3%
GPQA (HELM)	—	52%
MMLU	—	88.1%

Multimodal Not comparable

DeepSeek-V3.2-Exp: —, GPT-4o: 34.5 (#91)

Multimodal benchmarks
Benchmark	DeepSeek-V3.2-Exp	GPT-4o
LMArena Vision	—	1137
Video-MME	—	71.9%
GeoBench	—	71%
VPCT	—	40%
ScienceQA	—	88.5%

Multilingual DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 52.2 (#90), GPT-4o: 43.2 (#186)

Multilingual benchmarks
Benchmark	DeepSeek-V3.2-Exp	GPT-4o
LMArena Non-English	1409	1283
LMArena Chinese	1461	1277
LMArena French	1433	1304
LMArena German	1440	1282
LMArena Japanese	1374	1257
LMArena Korean	1371	1234
LMArena Russian	1424	1286
LMArena Spanish	1440	1292

Instruction Following DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 74.5 (#93), GPT-4o: 66.6 (#207)

Instruction Following benchmarks
Benchmark	DeepSeek-V3.2-Exp	GPT-4o
LMArena Instruction Following	1413	1278
LiveBench Instruction Following	—	68.6%
IFEval	—	81.7%

Long Context DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 47.6 (#16), GPT-4o: 39.4 (#179)

Long Context benchmarks
Benchmark	DeepSeek-V3.2-Exp	GPT-4o
Fiction.LiveBench	83.3%	66.7%
LMArena Longer Query	1428	1289
CL-bench	13.2%	—
CL-bench Life	9.5%	—

Writing & Preference DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 62.4 (#77), GPT-4o: 52.6 (#166)

Writing & Preference benchmarks
Benchmark	DeepSeek-V3.2-Exp	GPT-4o
LMArena Text	1425	1300
LMArena Creative Writing	1403	1292
LMArena Multi-Turn	1427	1302
Short-Story Creative Writing	—	81.8%
EQ-Bench Creative Writing	1515	—
WildBench	—	82.8%
LiveBench Language	—	47.6%

Frequently asked questions

Is DeepSeek-V3.2-Exp better than GPT-4o?

DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 28.6 on the Noometry Index.

Which is cheaper, DeepSeek-V3.2-Exp or GPT-4o?

DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; GPT-4o lists at $2.50 and $10.

Is DeepSeek-V3.2-Exp or GPT-4o better for coding?

DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 24.8 in the Noometry coding category.

Which has the bigger context window?

DeepSeek-V3.2-Exp does, with 164K tokens against 128K.

How many benchmarks do DeepSeek-V3.2-Exp and GPT-4o share?

33 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and GPT-4o has 72.