Model comparison

DeepSeek-V3.2-Exp vs Step 3

DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 40.5 on the Noometry Index.

Last verified October 10, 2026. 16 shared benchmarks.

DeepSeek-V3.2-Exp

DeepSeek

44.3

Rank #78 Confirmed

Step 3

StepFun

40.5

Rank #149 Confirmed

Summary

They share 16 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 7 categories and Step 3 in 1 category; 8 gaps are clear of the uncertainty.
The widest gap is in knowledge, where DeepSeek-V3.2-Exp leads 51.7 to 36.8.
The biggest single-benchmark swing is Kagi LLM Benchmark: 52.2% for DeepSeek-V3.2-Exp and 62.3% for Step 3.

Side by side

DeepSeek-V3.2-Exp and Step 3 specifications
	DeepSeek-V3.2-Exp	Step 3
Provider	DeepSeek	StepFun
Noometry Index	44.3	40.5
Released	2025-09-29	—
Weights	Open	Open
Context window	164K	—
Max output	66K	—
Input $ / M tokens	$0.26	—
Output $ / M tokens	$0.38	—
Results tracked	49	17

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), Step 3: 40.1 (#147)

Coding benchmarks
Benchmark	DeepSeek-V3.2-Exp	Step 3
LMArena Coding	1454	1367
SWE-bench Verified (bash only)	70%	—
Aider Polyglot	74.2%	—
LMArena WebDev	1362	—
SWE-bench Multilingual	59%	—
SciCode	38.9%	—
WeirdML	39.5%	—

Agentic & Tool Use Not comparable

DeepSeek-V3.2-Exp: 32.7 (#59), Step 3: —

Agentic & Tool Use benchmarks
Benchmark	DeepSeek-V3.2-Exp	Step 3
Terminal-Bench	39.6%	—
APEX-Agents	21.3%	—
Berkeley Function Calling Leaderboard	56.7%	—
TheAgentCompany	42.9%	—
Vending-Bench 2	1,034	—

Reasoning Step 3 leads

DeepSeek-V3.2-Exp: 22.1 (#208), Step 3: 28.4 (#105)

Reasoning benchmarks
Benchmark	DeepSeek-V3.2-Exp	Step 3
Kagi LLM Benchmark	52.2%	62.3%
LMArena Hard Prompts	1434	1355
ARC-AGI-2	4%	—
NYT Connections (extended)	36.7%	—
ARC-AGI-1	57%	—
CritPt	2.9%	—
Chess Puzzles	14%	—
Thematic Generalization	65%	—
DTBench	87.7%	—
LMCA	29.1%	—
Epoch Capabilities Index	146.27	—

Math DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 41.7 (#87), Step 3: 37.6 (#148)

Math benchmarks
Benchmark	DeepSeek-V3.2-Exp	Step 3
LMArena Math	1435	1366
MathArena Final-Answer Competitions	57.7%	—
OTIS Mock AIME 2024-2025	87.8%	—
ProofBench	8%	—
FrontierMath (Feb 2025 set)	22.1%	—
FrontierMath Tier 4 (v1)	2.1%	—

Knowledge DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 51.7 (#66), Step 3: 36.8 (#164)

Knowledge benchmarks
Benchmark	DeepSeek-V3.2-Exp	Step 3
LMArena Expert	1436	1333
GPQA Diamond	83.4%	—
Vectara Hallucination Rate	5.3%	—

Multimodal Not comparable

DeepSeek-V3.2-Exp: —, Step 3: 35.5 (#86)

Multimodal benchmarks
Benchmark	DeepSeek-V3.2-Exp	Step 3
LMArena Vision	—	1177

Multilingual DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 52.2 (#90), Step 3: 46.3 (#159)

Multilingual benchmarks
Benchmark	DeepSeek-V3.2-Exp	Step 3
LMArena Non-English	1409	1327
LMArena Chinese	1461	1397
LMArena German	1440	1371
LMArena Korean	1371	1269
LMArena Russian	1424	1331
LMArena Spanish	1440	1371
LMArena French	1433	—
LMArena Japanese	1374	—

Instruction Following DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 74.5 (#93), Step 3: 70.4 (#164)

Instruction Following benchmarks
Benchmark	DeepSeek-V3.2-Exp	Step 3
LMArena Instruction Following	1413	1332

Long Context DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 47.6 (#16), Step 3: 40.3 (#157)

Long Context benchmarks
Benchmark	DeepSeek-V3.2-Exp	Step 3
LMArena Longer Query	1428	1326
Fiction.LiveBench	83.3%	—
CL-bench	13.2%	—
CL-bench Life	9.5%	—

Writing & Preference DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 62.4 (#77), Step 3: 54.3 (#151)

Writing & Preference benchmarks
Benchmark	DeepSeek-V3.2-Exp	Step 3
LMArena Text	1425	1350
LMArena Creative Writing	1403	1321
LMArena Multi-Turn	1427	1341
EQ-Bench Creative Writing	1515	—

Frequently asked questions

Is DeepSeek-V3.2-Exp better than Step 3?