Enter opens the highlighted result. Esc closes.

DeepSeek V4.1 Flash vs Kimi K3

DeepSeek V4.1 Flash leads Kimi K3 on coding, 62.5 to 62.3 on the Vibe Coding Index, and DeepSeek V4.1 Flash costs 19.2x less per output token ($0.78 vs $15.00 per million). Updated October 21, 2026; 88 benchmarks have a result for both.

Replace DeepSeek V4.1 Flash Replace Kimi K3
DeepSeek V4.1 FlashDeepSeek
Coding
62.5
Overall
62.4
Output $/M
$0.78
14
Kimi K3Moonshot AI
Coding
62.3
Overall
66.5
Output $/M
$15.00
72
vs boards won of 88 shared, 2 tied

DeepSeek V4.1 Flash or Kimi K3: which to pick

DeepSeek V4.1 Flash and Kimi K3 are level for coding, 62.5 against 62.3 on the Vibe Coding Index, so price and context decide.

Pick Kimi K3 for work measured by LMArena WebDev or Vibe Code Bench (Vals AI), where it leads. Pick DeepSeek V4.1 Flash for output tokens 19.2x cheaper ($0.78 vs $15.00 per million) and work measured by LiveBench coding or Terminal-Bench 4 (Vals AI), where it leads.

Model ADeepSeek V4.1 FlashDeepSeek62.5/1005 independent benchmarksNo cost per solved taskModel BKimi K3Moonshot AI62.3/1006 independent benchmarks$0.81 per solved task
Shared results
88
DeepSeek V4.1 Flash only
13
Kimi K3 only
29
Like-for-like categories
8 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 V4.1 Flash 62.5 from 5 benchmarks, Kimi K3 62.3 from 6 benchmarks.
Widest category
Agentic work, 11.4 points apart on average over 10 shared benchmarks.
Widest single gap
LiveBench browser: DeepSeek V4.1 Flash 84.0% against Kimi K3 16.0%, 68.0 points apart on a 0 to 100 scale.
Output price
DeepSeek V4.1 Flash $0.78 against Kimi K3 $15.00 per million output tokens, 19.2x apart.
Context window
Both take 1.05M tokens.

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 V4.1 FlashBarsKimi K3Ties
Coding514
Intelligence02
Agentic work37
Reasoning361 tie
Math15
Knowledge07
Web design17
General1241 tie
All shared boards14722 ties
CodingLike-for-likeSix of these boards feed the Vibe Coding Index514
Coding benchmarks for DeepSeek V4.1 Flash and Kimi K3
BenchmarkDeepSeek V4.1 FlashKimi K3Leader
LiveBench coding90.2%59.3%DeepSeek V4.1 Flash
LMArena WebDev16181655Kimi K3
AA Coding IndexNot measured76.2Only Kimi K3 measured
LiveCodeBench (Vals AI)Not measured87.2%Only Kimi K3 measured
SWE-bench Verified (Vals AI)Not measured93.4%Only Kimi K3 measured
Terminal-Bench 4 (Vals AI)19.7%17.2%DeepSeek V4.1 Flash
Vibe Code Bench (Vals AI)84.7%85.0%Kimi K3
Code migration (Vals AI)45.6%16.1%DeepSeek V4.1 Flash
ProgramBench (Vals AI)0.5%2.0%Kimi K3
IOI (Vals AI)40.3%48.9%Kimi K3
Terminal-Bench 2.1 (Vals AI)74.5%80.9%Kimi K3
SWE-bench Multimodal (Vals AI)46.5%44.2%DeepSeek V4.1 Flash
Vibe Code Bench 1-100 (Vals AI)16.4%18.2%Kimi K3
SkillsBench (Vals AI)69.8%Not measuredOnly DeepSeek V4.1 Flash measured
CyberBench (Vals AI)73.7%75.2%Kimi K3
SciCode51.9%59.5%Kimi K3
FrontierCode43.9%44.2%Kimi K3
FrontierSWENot measured25.9%Only Kimi K3 measured
WeirdMLNot measured82.6%Only Kimi K3 measured
ALE-Bench1092.31524.5Kimi K3
LMArena Text: coding15021508Kimi K3
LMArena Text: software and IT14961499Kimi K3
LMArena WebDev: React16281670Kimi K3
LMArena WebDev: HTML16121644Kimi K3
LMArena WebDev: from an image16061579DeepSeek V4.1 Flash
Neither model has a result on the other 14 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for DeepSeek V4.1 Flash and Kimi K3
BenchmarkDeepSeek V4.1 FlashKimi K3Leader
AA Intelligence Index24.743.6Kimi K3
Epoch Capabilities Index155.9157.9Kimi K3
Agentic workLike-for-likePublic boards, shown as published, kept out of the index37
Agentic work benchmarks for DeepSeek V4.1 Flash and Kimi K3
BenchmarkDeepSeek V4.1 FlashKimi K3Leader
AA Agentic IndexNot measured50.0Only Kimi K3 measured
Design Arena: agents building web appsNot measured1295Only Kimi K3 measured
Design Arena: agents building full-stack appsNot measured1315Only Kimi K3 measured
Design Arena: agents building mobile appsNot measured1269Only Kimi K3 measured
Design Arena: agents building Android appsNot measured1250Only Kimi K3 measured
Design Arena: agents building Godot gamesNot measured1223Only Kimi K3 measured
Design Arena: agents building gamesNot measured1252Only Kimi K3 measured
Design Arena: HTML slidesNot measured1256Only Kimi K3 measured
Design Arena: PowerPoint slidesNot measured1273Only Kimi K3 measured
Design Arena: spreadsheetsNot measured1058Only Kimi K3 measured
LiveBench research86.8%79.0%DeepSeek V4.1 Flash
LiveBench browser84.0%16.0%DeepSeek V4.1 Flash
LiveBench long agentic40.1%42.8%Kimi K3
SREBench (Vals AI)0.8%Not measuredOnly DeepSeek V4.1 Flash measured
Finance Agent (Vals AI)53.5%53.1%DeepSeek V4.1 Flash
Tax Agent Bench (Vals AI)26.6%33.4%Kimi K3
Legal Research Bench (Vals AI)41.4%46.1%Kimi K3
Harvey's Legal Agent Benchmark (Vals AI)Not measured12.9%Only Kimi K3 measured
Excel Modeling Benchmark (Vals AI)57.2%66.7%Kimi K3
BioMysteryBench (Vals AI)67.8%72.6%Kimi K3
Terminal-Bench-Science (Vals AI)0.0%1.4%Kimi K3
Vals Science Index29.6%Not measuredOnly DeepSeek V4.1 Flash measured
Time Horizon Index (Vals AI)Not measured10.5%Only Kimi K3 measured
APEX-Agents39.5%47.8%Kimi K3
PostTrainBenchNot measured32.0%Only Kimi K3 measured
Vending-Bench 2Not measured$5,165Only Kimi K3 measured
Neither model has a result on the other 15 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index36
Reasoning benchmarks for DeepSeek V4.1 Flash and Kimi K3
BenchmarkDeepSeek V4.1 FlashKimi K3Leader
GPQA DiamondNot measured92.9%Only Kimi K3 measured
LiveBench reasoning86.7%90.7%Kimi K3
MysteryMechanism (Vals AI)21.2%Not measuredOnly DeepSeek V4.1 Flash measured
MMMU Pro (Vals AI)Not measured88.2%Only Kimi K3 measured
GPQA Diamond (Artificial Analysis run)Not measured93.5%Only Kimi K3 measured
Humanity's Last Exam (Artificial Analysis run)39.2%46.9%Kimi K3
GPQA Diamond (Epoch AI run)89.8%93.1%Kimi K3
ARC-AGI94.5%94.5%Tie
ARC-AGI-272.9%60.4%DeepSeek V4.1 Flash
SimpleBenchNot measured60.7%Only Kimi K3 measured
Chess puzzles48.0%37.0%DeepSeek V4.1 Flash
Mystery game puzzles43.0%26.0%DeepSeek V4.1 Flash
LMArena Text: hard prompts14831495Kimi K3
LMArena Text: expert15011520Kimi K3
LMArena Text: hard prompts in English14881496Kimi K3
LMArena Vision: diagrams1297Not measuredOnly DeepSeek V4.1 Flash measured
LMArena Vision: homework1307Not measuredOnly DeepSeek V4.1 Flash measured
Neither model has a result on the other 5 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index15
Math benchmarks for DeepSeek V4.1 Flash and Kimi K3
BenchmarkDeepSeek V4.1 FlashKimi K3Leader
OTIS Mock AIME98.3%97.2%DeepSeek V4.1 Flash
FrontierMath, tiers 1 to 367.4%72.2%Kimi K3
FrontierMath, tier 426.8%39.0%Kimi K3
ProofBench54.0%87.0%Kimi K3
LMArena Text: math14781494Kimi K3
LMArena Text: mathematical work14851505Kimi K3
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index07
Knowledge benchmarks for DeepSeek V4.1 Flash and Kimi K3
BenchmarkDeepSeek V4.1 FlashKimi K3Leader
MMLU ProNot measured88.0%Only Kimi K3 measured
LegalBench (Vals AI)83.3%86.2%Kimi K3
CorpFin (Vals AI)Not measured71.6%Only Kimi K3 measured
TaxEval (Vals AI)Not measured75.7%Only Kimi K3 measured
MortgageTax (Vals AI)Not measured66.3%Only Kimi K3 measured
MedCode (Vals AI)41.2%49.4%Kimi K3
SimpleQA Verified42.9%50.6%Kimi K3
LMArena Text: science14691508Kimi K3
LMArena Text: medicine and healthcare14841489Kimi K3
LMArena Text: legal and government14431503Kimi K3
LMArena Text: business and finance14611470Kimi K3
Neither model has a result on the other 7 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index17
Web design benchmarks for DeepSeek V4.1 Flash and Kimi K3
BenchmarkDeepSeek V4.1 FlashKimi K3Leader
Design Arena: websites13091335Kimi K3
Design Arena: UI components13151344Kimi K3
Design Arena: game dev13301359Kimi K3
Design Arena: data viz12901351Kimi K3
Design Arena: 3D13561388Kimi K3
Design Arena: all code categories13201360Kimi K3
Design Arena: SVG12441302Kimi K3
Design Arena: ASCII art1265Not measuredOnly DeepSeek V4.1 Flash measured
LiveBench design79.2%77.2%DeepSeek V4.1 Flash
GeneralLike-for-likePublic boards, shown as published, kept out of the index124
General benchmarks for DeepSeek V4.1 Flash and Kimi K3
BenchmarkDeepSeek V4.1 FlashKimi K3Leader
Vals IndexNot measured53.0%Only Kimi K3 measured
LiveBench data analysis79.3%78.7%DeepSeek V4.1 Flash
Public Benefits Bench (Vals AI)64.3%68.2%Kimi K3
SAGE (Vals AI)47.9%52.8%Kimi K3
MedScribe (Vals AI)85.5%88.0%Kimi K3
Vals Multimodal IndexNot measured73.4%Only Kimi K3 measured
Long context reasoning (AA-LCR)84.0%88.7%Kimi K3
DTBench89.9%91.2%Kimi K3
LMCA47.0%52.7%Kimi K3
Surface Evolver Bench46.3%95.0%Kimi K3
Furniture assembly34.2%34.2%Tie
LMArena Text14641476Kimi K3
LMArena Text: instruction following14731485Kimi K3
LMArena Text: creative writing14371457Kimi K3
LMArena Vision1278Not measuredOnly DeepSeek V4.1 Flash measured
LMArena Text: English14731481Kimi K3
LMArena Text: non-English14491466Kimi K3
LMArena Text: Chinese15011529Kimi K3
LMArena Text: German14701490Kimi K3
LMArena Text: Japanese14321494Kimi K3
LMArena Text: Korean14361460Kimi K3
LMArena Text: French14651492Kimi K3
LMArena Text: Spanish14591472Kimi K3
LMArena Text: Russian14671485Kimi K3
LMArena Text: Polish14571458Kimi K3
LMArena Text: multi-turn14611489Kimi K3
LMArena Text: longer queries14731494Kimi K3
LMArena Text: writing and language14501471Kimi K3
LMArena Text: entertainment and media14451453Kimi K3
LMArena Vision: English1288Not measuredOnly DeepSeek V4.1 Flash measured
LMArena Vision: Chinese1327Not measuredOnly DeepSeek V4.1 Flash measured
LMArena Vision: reading text1282Not measuredOnly DeepSeek V4.1 Flash measured
LMArena Vision: humor1260Not measuredOnly DeepSeek V4.1 Flash measured
LMArena Vision: creative writing1275Not measuredOnly DeepSeek V4.1 Flash measured
Neither model has a result on the other 12 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for DeepSeek V4.1 Flash and Kimi K3
BenchmarkDeepSeek V4.1 FlashKimi K3Leader
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 V4.1 Flash$0.0006Fits in one request
Kimi K3$0.011Fits in one request

DeepSeek V4.1 Flash costs less

Repository review

50K input + 3K output tokens

DeepSeek V4.1 Flash$0.013Fits in one request
Kimi K3$0.195Fits in one request

DeepSeek V4.1 Flash costs less

Agent loop

200K input + 10K output tokens

DeepSeek V4.1 Flash$0.050Fits in one request
Kimi K3$0.750Fits in one request

DeepSeek V4.1 Flash 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 V4.1 Flash and Kimi K3
SpecificationDeepSeek V4.1 FlashMoonshotAI Kimi K3
MakerDeepSeekMoonshot AI
WeightsOpen weightsOpen weights
ReleasedSeptember 10, 2026July 16, 2026
Context window1.05M1.05M
Input $/Mtok$0.21$3.00
Output $/Mtok$0.78$15.00
Coding index62.562.3
Independent measurements behind it56
Cost per solved taskNot computable$0.81
SourceOpenRouter listingOpenRouter listing

Questions

Is DeepSeek V4.1 Flash better than Kimi K3 for coding?

DeepSeek V4.1 Flash and Kimi K3 are level for coding, 62.5 against 62.3 on the Vibe Coding Index, so price and context decide. Pick Kimi K3 for work measured by LMArena WebDev or Vibe Code Bench (Vals AI), where it leads. Pick DeepSeek V4.1 Flash for output tokens 19.2x cheaper ($0.78 vs $15.00 per million) and work measured by LiveBench coding or Terminal-Bench 4 (Vals AI), where it leads.

Which is cheaper, DeepSeek V4.1 Flash or Kimi K3?

DeepSeek V4.1 Flash costs $0.21 per million input tokens and $0.78 per million output tokens; Kimi K3 costs $3.00 and $15.00. Cost per solved task is known for Kimi K3 only.

Which has the bigger context window, DeepSeek V4.1 Flash or Kimi K3?

Neither: both take 1.05M tokens.

How many benchmarks is the DeepSeek V4.1 Flash vs Kimi K3 comparison based on?

88 benchmarks have a result for both models as of October 21, 2026. DeepSeek V4.1 Flash's coding index rests on 5 independent benchmarks and Kimi K3's on 6. 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-v4-1-flash-vs-kimi-k3" title="DeepSeek V4.1 Flash vs Kimi K3, 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-v4-1-flash-vs-kimi-k3">LLMBenchmark.ai, DeepSeek V4.1 Flash vs Kimi K3</a></p>