Enter opens the highlighted result. Esc closes.

DeepSeek V3 0324 vs DeepSeek V4 Flash 0731

DeepSeek V4 Flash 0731 leads DeepSeek V3 0324 on the Artificial Analysis Intelligence Index, 34.3 to 9.7, and DeepSeek V4 Flash 0731 costs 2.7x less per output token ($0.396 vs $1.07 per million). Updated October 22, 2026; 19 benchmarks have a result for both.

Replace DeepSeek V3 0324 Replace DeepSeek V4 Flash 0731
DeepSeek V3 0324DeepSeek
Coding
No data
Overall
No data
Output $/M
$1.07
2
DeepSeek V4 Flash 0731DeepSeek
Coding
51.8
Overall
55.2
Output $/M
$0.396
17
vs boards won of 19 shared

DeepSeek V3 0324 or DeepSeek V4 Flash 0731: which to pick

Only DeepSeek V4 Flash 0731 has a coding score, 51.8 on the Vibe Coding Index; DeepSeek V3 0324 has none, so the two cannot be ranked on coding.

Pick DeepSeek V3 0324 for work measured by LegalBench (Vals AI) or TaxEval (Vals AI), where it leads.

Model ADeepSeek V3 0324DeepSeekNot benchmarkedNo coding benchmarkNo cost per solved taskModel BDeepSeek V4 Flash 0731DeepSeek51.8/1004 independent benchmarks$0.01 per solved task
Shared results
19
DeepSeek V3 0324 only
16
DeepSeek V4 Flash 0731 only
36
Like-for-like categories
7 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

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 0324 not benchmarked, DeepSeek V4 Flash 0731 51.8 from 4 benchmarks.
Widest category
Math, 56.7 points apart on average over 1 shared benchmark.
Widest single gap
OTIS Mock AIME: DeepSeek V3 0324 37.8% against DeepSeek V4 Flash 0731 94.4%, 56.7 points apart on a 0 to 100 scale.
Output price
DeepSeek V3 0324 $1.07 against DeepSeek V4 Flash 0731 $0.396 per million output tokens, 2.7x apart.
Context window
DeepSeek V3 0324 takes 164K tokens, DeepSeek V4 Flash 0731 1.05M.

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 0324BarsDeepSeek V4 Flash 0731Ties
Coding04
Intelligence01
Agentic work01
Reasoning05
Math01
Knowledge22
General03
All shared boards217
CodingLike-for-likeSix of these boards feed the Vibe Coding Index04
Coding benchmarks for DeepSeek V3 0324 and DeepSeek V4 Flash 0731
BenchmarkDeepSeek V3 0324DeepSeek V4 Flash 0731Leader
SWE-bench Verified42.0%Not measuredOnly DeepSeek V3 0324 measured
AA Coding Index21.269.1DeepSeek V4 Flash 0731
LiveCodeBench (Vals AI)65.5%87.3%DeepSeek V4 Flash 0731
SWE-bench Verified (Vals AI)Not measured88.8%Only DeepSeek V4 Flash 0731 measured
Terminal-Bench 4 (Vals AI)Not measured18.7%Only DeepSeek V4 Flash 0731 measured
Vibe Code Bench (Vals AI)Not measured74.7%Only DeepSeek V4 Flash 0731 measured
Code migration (Vals AI)Not measured38.6%Only DeepSeek V4 Flash 0731 measured
ProgramBench (Vals AI)Not measured0.0%Only DeepSeek V4 Flash 0731 measured
IOI (Vals AI)Not measured32.7%Only DeepSeek V4 Flash 0731 measured
Terminal-Bench 2.1 (Vals AI)Not measured67.0%Only DeepSeek V4 Flash 0731 measured
SkillsBench (Vals AI)Not measured50.7%Only DeepSeek V4 Flash 0731 measured
CyberBench (Vals AI)Not measured72.1%Only DeepSeek V4 Flash 0731 measured
LiveCodeBench (Artificial Analysis run)40.5%Not measuredOnly DeepSeek V3 0324 measured
SciCode39.0%50.3%DeepSeek V4 Flash 0731
Terminal-Bench Hard (Artificial Analysis run)15.2%Not measuredOnly DeepSeek V3 0324 measured
FrontierCodeNot measured18.8%Only DeepSeek V4 Flash 0731 measured
WeirdML36.1%63.0%DeepSeek V4 Flash 0731
ALE-BenchNot measured1306.1Only DeepSeek V4 Flash 0731 measured
Neither model has a result on the other 21 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
Intelligence benchmarks for DeepSeek V3 0324 and DeepSeek V4 Flash 0731
BenchmarkDeepSeek V3 0324DeepSeek V4 Flash 0731Leader
AA Intelligence Index9.734.3DeepSeek V4 Flash 0731
Epoch Capabilities IndexNot measured154.8Only DeepSeek V4 Flash 0731 measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for DeepSeek V3 0324 and DeepSeek V4 Flash 0731
BenchmarkDeepSeek V3 0324DeepSeek V4 Flash 0731Leader
AA Agentic Index0.841.0DeepSeek V4 Flash 0731
Finance Agent (Vals AI)Not measured49.5%Only DeepSeek V4 Flash 0731 measured
Tax Agent Bench (Vals AI)Not measured28.1%Only DeepSeek V4 Flash 0731 measured
Legal Research Bench (Vals AI)Not measured30.3%Only DeepSeek V4 Flash 0731 measured
Excel Modeling Benchmark (Vals AI)Not measured57.0%Only DeepSeek V4 Flash 0731 measured
BioMysteryBench (Vals AI)Not measured64.4%Only DeepSeek V4 Flash 0731 measured
Terminal-Bench-Science (Vals AI)Not measured4.3%Only DeepSeek V4 Flash 0731 measured
τ²-Bench Telecom47.1%Not measuredOnly DeepSeek V3 0324 measured
METR time horizons49.6%Not measuredOnly DeepSeek V3 0324 measured
Neither model has a result on the other 32 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index05
Reasoning benchmarks for DeepSeek V3 0324 and DeepSeek V4 Flash 0731
BenchmarkDeepSeek V3 0324DeepSeek V4 Flash 0731Leader
GPQA Diamond61.6%89.9%DeepSeek V4 Flash 0731
GPQA Diamond (Artificial Analysis run)65.5%90.8%DeepSeek V4 Flash 0731
Humanity's Last Exam (Artificial Analysis run)4.7%38.6%DeepSeek V4 Flash 0731
GPQA Diamond (Epoch AI run)67.6%91.0%DeepSeek V4 Flash 0731
ARC-AGINot measured89.0%Only DeepSeek V4 Flash 0731 measured
ARC-AGI-2Not measured61.4%Only DeepSeek V4 Flash 0731 measured
SimpleBench27.2%61.1%DeepSeek V4 Flash 0731
Chess puzzlesNot measured32.0%Only DeepSeek V4 Flash 0731 measured
Mystery game puzzlesNot measured34.0%Only DeepSeek V4 Flash 0731 measured
Neither model has a result on the other 13 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index01
Math benchmarks for DeepSeek V3 0324 and DeepSeek V4 Flash 0731
BenchmarkDeepSeek V3 0324DeepSeek V4 Flash 0731Leader
AIME52.2%Not measuredOnly DeepSeek V3 0324 measured
MATH 500 (Vals AI)88.6%Not measuredOnly DeepSeek V3 0324 measured
MGSM (Vals AI)91.7%Not measuredOnly DeepSeek V3 0324 measured
AA Math Index41.0Not measuredOnly DeepSeek V3 0324 measured
AIME 2025 (Artificial Analysis run)41.0%Not measuredOnly DeepSeek V3 0324 measured
OTIS Mock AIME37.8%94.4%DeepSeek V4 Flash 0731
FrontierMath, tiers 1 to 3Not measured57.5%Only DeepSeek V4 Flash 0731 measured
FrontierMath, tier 4Not measured24.4%Only DeepSeek V4 Flash 0731 measured
ProofBenchNot measured56.0%Only DeepSeek V4 Flash 0731 measured
MATH level 575.5%Not measuredOnly DeepSeek V3 0324 measured
Neither model has a result on the other 3 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index22
Knowledge benchmarks for DeepSeek V3 0324 and DeepSeek V4 Flash 0731
BenchmarkDeepSeek V3 0324DeepSeek V4 Flash 0731Leader
MMLU Pro79.5%86.2%DeepSeek V4 Flash 0731
LegalBench (Vals AI)77.7%77.7%DeepSeek V3 0324
CorpFin (Vals AI)54.7%61.9%DeepSeek V4 Flash 0731
TaxEval (Vals AI)71.1%70.7%DeepSeek V3 0324
MedQA (Vals AI)82.0%Not measuredOnly DeepSeek V3 0324 measured
MedCode (Vals AI)Not measured41.4%Only DeepSeek V4 Flash 0731 measured
MMLU Pro (Artificial Analysis run)81.9%Not measuredOnly DeepSeek V3 0324 measured
SimpleQA VerifiedNot measured33.6%Only DeepSeek V4 Flash 0731 measured
Neither model has a result on the other 10 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
Web design benchmarks for DeepSeek V3 0324 and DeepSeek V4 Flash 0731
BenchmarkDeepSeek V3 0324DeepSeek V4 Flash 0731Leader
Design Arena: websitesNot measured1237Only DeepSeek V4 Flash 0731 measured
Design Arena: UI componentsNot measured1233Only DeepSeek V4 Flash 0731 measured
Design Arena: game devNot measured1207Only DeepSeek V4 Flash 0731 measured
Design Arena: data vizNot measured1180Only DeepSeek V4 Flash 0731 measured
Design Arena: 3DNot measured1208Only DeepSeek V4 Flash 0731 measured
Design Arena: all code categoriesNot measured1229Only DeepSeek V4 Flash 0731 measured
Design Arena: SVGNot measured1184Only DeepSeek V4 Flash 0731 measured
Design Arena: ASCII artNot measured1090Only DeepSeek V4 Flash 0731 measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index03
General benchmarks for DeepSeek V3 0324 and DeepSeek V4 Flash 0731
BenchmarkDeepSeek V3 0324DeepSeek V4 Flash 0731Leader
MedScribe (Vals AI)Not measured80.4%Only DeepSeek V4 Flash 0731 measured
IFBench41.0%Not measuredOnly DeepSeek V3 0324 measured
Long context reasoning (AA-LCR)40.7%79.7%DeepSeek V4 Flash 0731
DTBench64.8%90.9%DeepSeek V4 Flash 0731
LMCA15.5%41.7%DeepSeek V4 Flash 0731
Fiction.LiveBench50.0%Not measuredOnly DeepSeek V3 0324 measured
Creative writing (Lech Mazur)77.0%Not measuredOnly DeepSeek V3 0324 measured
Neither model has a result on the other 39 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for DeepSeek V3 0324 and DeepSeek V4 Flash 0731
BenchmarkDeepSeek V3 0324DeepSeek V4 Flash 0731Leader
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 0324$0.0008Fits in one request
DeepSeek V4 Flash 0731$0.0003Fits in one request

DeepSeek V4 Flash 0731 costs less

Repository review

50K input + 3K output tokens

DeepSeek V3 0324$0.017Fits in one request
DeepSeek V4 Flash 0731$0.0082Fits in one request

DeepSeek V4 Flash 0731 costs less

Agent loop

200K input + 10K output tokens

DeepSeek V3 0324$0.065Exceeds the context window
DeepSeek V4 Flash 0731$0.032Fits in one request

DeepSeek V4 Flash 0731 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 0324 and DeepSeek V4 Flash 0731
SpecificationDeepSeek V3 0324DeepSeek V4 Flash 0731
MakerDeepSeekDeepSeek
WeightsOpen weightsOpen weights
ReleasedMarch 24, 2025July 31, 2026
Context window164K1.05M
Input $/Mtok$0.27$0.14
Output $/Mtok$1.07$0.396
Coding indexNot benchmarked51.8
Independent measurements behind it04
Cost per solved taskNot computable$0.01
SourceOpenRouter listingOpenRouter listing

Questions

Is DeepSeek V3 0324 better than DeepSeek V4 Flash 0731 for coding?

Only DeepSeek V4 Flash 0731 has a coding score, 51.8 on the Vibe Coding Index; DeepSeek V3 0324 has none, so the two cannot be ranked on coding. Pick DeepSeek V3 0324 for work measured by LegalBench (Vals AI) or TaxEval (Vals AI), where it leads.

Which is cheaper, DeepSeek V3 0324 or DeepSeek V4 Flash 0731?

DeepSeek V3 0324 costs $0.27 per million input tokens and $1.07 per million output tokens; DeepSeek V4 Flash 0731 costs $0.14 and $0.396. Cost per solved task is known for DeepSeek V4 Flash 0731 only.

Which has the bigger context window, DeepSeek V3 0324 or DeepSeek V4 Flash 0731?

DeepSeek V4 Flash 0731, with 1.05M tokens against 164K.

How many benchmarks is the DeepSeek V3 0324 vs DeepSeek V4 Flash 0731 comparison based on?

19 benchmarks have a result for both models as of October 22, 2026. DeepSeek V3 0324's coding index rests on 0 independent benchmarks and DeepSeek V4 Flash 0731's on 4. Where only one model has a result, the page says so instead of naming a winner.

Compare next