LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

DeepSeek V3.2 vs o4 Mini

Updated October 21, 2026. 60 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace DeepSeek V3.2 Replace o4 Mini
DeepSeek V3.2DeepSeek
Coding
54.4
Overall
21.5
Output $/M
$0.42
51
o4 MiniOpenAI
Coding
68.6
Overall
16.7
Output $/M
$4.40
9
vs boards won of 60 shared

Decision reading

Pick o4 Mini for coding: it scores 68.6 on the Vibe Coding Index against 54.4 for DeepSeek V3.2.

Pick DeepSeek V3.2 when output tokens cost 10.5x less, or when what matters is SWE-bench Verified or LiveCodeBench (Artificial Analysis run), where it leads.

Model ADeepSeek V3.2DeepSeek54.4/1002 independent benchmarks$0.99 per solved taskModel Bo4 MiniOpenAI68.6/1002 independent benchmarksNo cost per solved task
Shared results
60
DeepSeek V3.2 only
15
o4 Mini only
34
Like-for-like categories
8 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
DeepSeek V3.2 54.4 from 2 benchmarks, o4 Mini 68.6 from 2 benchmarks.
Widest category
Agentic work, 35.1 points apart on average over 1 shared benchmark.
Widest single gap
τ²-Bench Telecom: DeepSeek V3.2 90.6% against o4 Mini 55.6%, 35.1 points apart on a 0 to 100 scale.
Output price
DeepSeek V3.2 $0.42 against o4 Mini $4.40 per million output tokens, 10.5x apart.
Context window
DeepSeek V3.2 takes 164K tokens, o4 Mini 200K.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryDeepSeek V3.2Barso4 MiniTies
Coding51
Intelligence20
Agentic work10
Reasoning62
Math51
Knowledge64
Web design60
General201
All shared boards519
CodingLike-for-likeFour of these boards feed the Vibe Coding Index51
Coding benchmarks for DeepSeek V3.2 and o4 Mini
BenchmarkDeepSeek V3.2o4 MiniLeader
Aider polyglotNot measured72.0%Only o4 Mini measured
SWE-bench Verified70.0%64.6%DeepSeek V3.2
SWE-bench MultimodalNot measured33.9%Only o4 Mini measured
SWE-bench Multilingual59.0%Not measuredOnly DeepSeek V3.2 measured
LMArena WebDev1362Not measuredOnly DeepSeek V3.2 measured
AA Coding Index44.2Not measuredOnly DeepSeek V3.2 measured
LiveCodeBench (Vals AI)80.7%82.2%o4 Mini
SWE-bench Verified (Vals AI)67.6%Not measuredOnly DeepSeek V3.2 measured
Vibe Code Bench (Vals AI)5.1%Not measuredOnly DeepSeek V3.2 measured
LiveCodeBench (Artificial Analysis run)86.2%85.9%DeepSeek V3.2
Terminal-Bench Hard (Artificial Analysis run)35.6%15.2%DeepSeek V3.2
Terminal-Bench39.6%Not measuredOnly DeepSeek V3.2 measured
GSO-BenchNot measured3.6%Only o4 Mini measured
WeirdMLNot measured52.6%Only o4 Mini measured
LMArena Text: coding14531368DeepSeek V3.2
LMArena Text: software and IT14491367DeepSeek V3.2
LMArena WebDev: React1350Not measuredOnly DeepSeek V3.2 measured
LMArena WebDev: HTML1371Not measuredOnly DeepSeek V3.2 measured
Neither model has a result on the other 20 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for DeepSeek V3.2 and o4 Mini
BenchmarkDeepSeek V3.2o4 MiniLeader
AA Intelligence Index21.516.7DeepSeek V3.2
Epoch Capabilities Index146.1145.5DeepSeek V3.2
Agentic workLike-for-likePublic boards, shown as published, kept out of the index10
Agentic work benchmarks for DeepSeek V3.2 and o4 Mini
BenchmarkDeepSeek V3.2o4 MiniLeader
τ²-Bench Telecom90.6%55.6%DeepSeek V3.2
APEX-Agents21.3%Not measuredOnly DeepSeek V3.2 measured
METR time horizonsNot measured63.9%Only o4 Mini measured
Vending-Bench 2$1,034Not measuredOnly DeepSeek V3.2 measured
Neither model has a result on the other 35 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index62
Reasoning benchmarks for DeepSeek V3.2 and o4 Mini
BenchmarkDeepSeek V3.2o4 MiniLeader
GPQA Diamond80.3%74.5%DeepSeek V3.2
MMMU Pro (Vals AI)Not measured79.7%Only o4 Mini measured
GPQA Diamond (Artificial Analysis run)84.0%78.4%DeepSeek V3.2
Humanity's Last Exam (Artificial Analysis run)24.6%16.5%DeepSeek V3.2
GPQA Diamond (Epoch AI run)Not measured79.6%Only o4 Mini measured
ARC-AGI57.0%58.7%o4 Mini
ARC-AGI-24.0%6.1%o4 Mini
SimpleBenchNot measured38.7%Only o4 Mini measured
Chess puzzlesNot measured22.0%Only o4 Mini measured
Mystery game puzzlesNot measured5.0%Only o4 Mini measured
VPCTNot measured57.5%Only o4 Mini measured
LMArena Text: hard prompts14341351DeepSeek V3.2
LMArena Text: expert14361343DeepSeek V3.2
LMArena Text: hard prompts in English14461366DeepSeek V3.2
LMArena Vision: diagramsNot measured1197Only o4 Mini measured
LMArena Vision: homeworkNot measured1234Only o4 Mini measured
Neither model has a result on the other 5 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index51
Math benchmarks for DeepSeek V3.2 and o4 Mini
BenchmarkDeepSeek V3.2o4 MiniLeader
AIME84.6%83.7%DeepSeek V3.2
MATH 500 (Vals AI)Not measured94.2%Only o4 Mini measured
MGSM (Vals AI)86.0%93.4%o4 Mini
AA Math Index92.090.7DeepSeek V3.2
AIME 2025 (Artificial Analysis run)92.0%90.7%DeepSeek V3.2
OTIS Mock AIMENot measured81.7%Only o4 Mini measured
FrontierMath, tiers 1 to 3Not measured36.1%Only o4 Mini measured
FrontierMath, tier 4Not measured4.9%Only o4 Mini measured
MATH level 5Not measured97.8%Only o4 Mini measured
LMArena Text: math14351389DeepSeek V3.2
LMArena Text: mathematical work14401379DeepSeek V3.2
Neither model has a result on the other 2 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index64
Knowledge benchmarks for DeepSeek V3.2 and o4 Mini
BenchmarkDeepSeek V3.2o4 MiniLeader
MMLU Pro84.9%80.6%DeepSeek V3.2
LegalBench (Vals AI)76.1%79.2%o4 Mini
CaseLaw (Vals AI)55.4%Not measuredOnly DeepSeek V3.2 measured
CorpFin (Vals AI)51.0%59.0%o4 Mini
TaxEval (Vals AI)68.2%74.8%o4 Mini
MortgageTax (Vals AI)Not measured64.8%Only o4 Mini measured
MedQA (Vals AI)93.9%96.0%o4 Mini
MedCode (Vals AI)Not measured33.8%Only o4 Mini measured
MMLU Pro (Artificial Analysis run)86.2%83.2%DeepSeek V3.2
SimpleQA VerifiedNot measured19.6%Only o4 Mini measured
GeoBenchNot measured64.0%Only o4 Mini measured
LMArena Text: science14461354DeepSeek V3.2
LMArena Text: medicine and healthcare14401361DeepSeek V3.2
LMArena Text: legal and government14251350DeepSeek V3.2
LMArena Text: business and finance14101325DeepSeek V3.2
Neither model has a result on the other 2 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index60
Web design benchmarks for DeepSeek V3.2 and o4 Mini
BenchmarkDeepSeek V3.2o4 MiniLeader
Design Arena: websites1175985DeepSeek V3.2
Design Arena: UI components1149983DeepSeek V3.2
Design Arena: game dev11321004DeepSeek V3.2
Design Arena: data viz1158990DeepSeek V3.2
Design Arena: 3D1133854DeepSeek V3.2
Design Arena: all code categories1164977DeepSeek V3.2
Design Arena: SVG1032Not measuredOnly DeepSeek V3.2 measured
Design Arena: ASCII art1082Not measuredOnly DeepSeek V3.2 measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index201
General benchmarks for DeepSeek V3.2 and o4 Mini
BenchmarkDeepSeek V3.2o4 MiniLeader
SAGE (Vals AI)Not measured41.1%Only o4 Mini measured
MedScribe (Vals AI)Not measured69.1%Only o4 Mini measured
IFBench60.7%68.7%o4 Mini
Long context reasoning (AA-LCR)73.3%61.0%DeepSeek V3.2
DTBench85.6%77.6%DeepSeek V3.2
LMCA28.8%26.5%DeepSeek V3.2
Fiction.LiveBenchNot measured77.8%Only o4 Mini measured
Creative writing (Lech Mazur)Not measured75.0%Only o4 Mini measured
CL-bench12.4%Not measuredOnly DeepSeek V3.2 measured
CL-bench Life7.4%Not measuredOnly DeepSeek V3.2 measured
LMArena Text14251353DeepSeek V3.2
LMArena Text: instruction following14131320DeepSeek V3.2
LMArena Text: creative writing14001294DeepSeek V3.2
LMArena VisionNot measured1194Only o4 Mini measured
LMArena Text: English14391367DeepSeek V3.2
LMArena Text: non-English14091337DeepSeek V3.2
LMArena Text: Chinese14551354DeepSeek V3.2
LMArena Text: German14131336DeepSeek V3.2
LMArena Text: Japanese13741308DeepSeek V3.2
LMArena Text: Korean13701312DeepSeek V3.2
LMArena Text: French14291364DeepSeek V3.2
LMArena Text: Spanish14241346DeepSeek V3.2
LMArena Text: Russian14241334DeepSeek V3.2
LMArena Text: Polish14191358DeepSeek V3.2
LMArena Text: multi-turn14271350DeepSeek V3.2
LMArena Text: longer queries14281315DeepSeek V3.2
LMArena Text: writing and language14071316DeepSeek V3.2
LMArena Text: entertainment and media13951308DeepSeek V3.2
LMArena Vision: EnglishNot measured1200Only o4 Mini measured
LMArena Vision: ChineseNot measured1218Only o4 Mini measured
LMArena Vision: reading textNot measured1197Only o4 Mini measured
LMArena Vision: humorNot measured1199Only o4 Mini measured
LMArena Vision: captioningNot measured1204Only o4 Mini measured
LMArena Vision: recognizing thingsNot measured1218Only o4 Mini measured
LMArena Vision: creative writingNot measured1166Only o4 Mini measured
Neither model has a result on the other 11 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for DeepSeek V3.2 and o4 Mini
BenchmarkDeepSeek V3.2o4 MiniLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

DeepSeek V3.2$0.0005Fits in one request
o4 Mini$0.0033Fits in one request

DeepSeek V3.2 costs less

Repository review

50K input + 3K output tokens

DeepSeek V3.2$0.015Fits in one request
o4 Mini$0.068Fits in one request

DeepSeek V3.2 costs less

Agent loop

200K input + 10K output tokens

DeepSeek V3.2$0.060Exceeds the context window
o4 Mini$0.264Exceeds the context window

DeepSeek V3.2 costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of DeepSeek V3.2 and o4 Mini
SpecificationDeepSeek V3.2OpenAI o4 Mini
MakerDeepSeekOpenAI
WeightsOpen weightsNo public weights listed
ReleasedDecember 1, 2025April 16, 2025
Context window164K200K
Input $/Mtok$0.28$1.10
Output $/Mtok$0.42$4.40
Coding index54.468.6
Independent measurements behind it22
Cost per solved task$0.99Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is DeepSeek V3.2 better than o4 Mini for coding?

Pick o4 Mini for coding: it scores 68.6 on the Vibe Coding Index against 54.4 for DeepSeek V3.2. Pick DeepSeek V3.2 when output tokens cost 10.5x less, or when what matters is SWE-bench Verified or LiveCodeBench (Artificial Analysis run), where it leads.

Which is cheaper, DeepSeek V3.2 or o4 Mini?

DeepSeek V3.2 costs $0.28 per million input tokens and $0.42 per million output tokens; o4 Mini costs $1.10 and $4.40. Cost per solved task is known for DeepSeek V3.2 only.

Which has the bigger context window, DeepSeek V3.2 or o4 Mini?

o4 Mini, with 200K tokens against 164K.

How many benchmarks is the DeepSeek V3.2 vs o4 Mini comparison based on?

60 benchmarks have a result for both models as of October 21, 2026. DeepSeek V3.2's coding index rests on 2 independent benchmarks and o4 Mini's on 2. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/deepseek-v3-2-vs-o4-mini" title="DeepSeek V3.2 vs o4 Mini, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/deepseek-v3-2-vs-o4-mini">LLMBenchmark.ai, DeepSeek V3.2 vs o4 Mini</a></p>