Enter opens the highlighted result. Esc closes.

Kimi K2.6 vs Kimi K3

Kimi K3 leads Kimi K2.6 on coding, 62.3 to 52.1 on the Vibe Coding Index, and Kimi K2.6 costs 3.8x less per output token ($4.00 vs $15.00 per million). Updated October 22, 2026; 92 benchmarks have a result for both.

Replace Kimi K2.6 Replace Kimi K3
Kimi K2.6Moonshot AI
Coding
52.1
Overall
60.5
Output $/M
$4.00
3
Kimi K3Moonshot AI
Coding
62.3
Overall
66.5
Output $/M
$15.00
89
vs boards won of 92 shared

Kimi K2.6 or Kimi K3: which to pick

Pick Kimi K3 for coding: it scores 62.3 on the Vibe Coding Index against 52.1 for Kimi K2.6.

Pick Kimi K2.6 for output tokens 3.8x cheaper ($4.00 vs $15.00 per million) and a lower cost per solved task ($0.35 against $0.81).

Model AKimi K2.6Moonshot AI52.1/1003 independent benchmarks$0.35 per solved taskModel BKimi K3Moonshot AI62.3/1006 independent benchmarks$0.81 per solved task
Shared results
92
Kimi K2.6 only
17
Kimi K3 only
25
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
Kimi K2.6 52.1 from 3 benchmarks, Kimi K3 62.3 from 6 benchmarks.
Widest category
Math, 18.0 points apart on average over 6 shared benchmarks.
Widest single gap
ProofBench: Kimi K2.6 16.0% against Kimi K3 87.0%, 71.0 points apart on a 0 to 100 scale.
Output price
Kimi K2.6 $4.00 against Kimi K3 $15.00 per million output tokens, 3.8x apart.
Context window
Kimi K2.6 takes 262K tokens, Kimi K3 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
CategoryKimi K2.6BarsKimi K3Ties
Coding115
Intelligence02
Agentic work213
Reasoning010
Math06
Knowledge011
Web design07
General025
All shared boards389
CodingLike-for-likeSix of these boards feed the Vibe Coding Index115
Coding benchmarks for Kimi K2.6 and Kimi K3
BenchmarkKimi K2.6Kimi K3Leader
LiveBench codingNot measured59.3%Only Kimi K3 measured
LMArena WebDev15091655Kimi K3
AA Coding Index61.876.2Kimi K3
LiveCodeBench (Vals AI)86.8%87.2%Kimi K3
SWE-bench Verified (Vals AI)76.2%93.4%Kimi K3
Terminal-Bench 4 (Vals AI)Not measured17.2%Only Kimi K3 measured
Vibe Code Bench (Vals AI)37.9%85.0%Kimi K3
Code migration (Vals AI)27.8%16.1%Kimi K2.6
ProgramBench (Vals AI)0.0%2.0%Kimi K3
IOI (Vals AI)Not measured48.9%Only Kimi K3 measured
Terminal-Bench 2.1 (Vals AI)53.6%80.9%Kimi K3
SWE-bench Multimodal (Vals AI)Not measured44.2%Only Kimi K3 measured
Vibe Code Bench 1-100 (Vals AI)Not measured18.2%Only Kimi K3 measured
CyberBench (Vals AI)Not measured75.2%Only Kimi K3 measured
SciCode51.5%59.5%Kimi K3
Terminal-Bench Hard (Artificial Analysis run)43.9%Not measuredOnly Kimi K2.6 measured
FrontierCodeNot measured44.2%Only Kimi K3 measured
FrontierSWENot measured25.9%Only Kimi K3 measured
WeirdML55.9%82.6%Kimi K3
ALE-Bench1092.71524.5Kimi K3
LMArena Text: coding14891508Kimi K3
LMArena Text: software and IT14841499Kimi K3
LMArena WebDev: React15051670Kimi K3
LMArena WebDev: HTML14851644Kimi K3
LMArena WebDev: from an image15211579Kimi K3
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 Kimi K2.6 and Kimi K3
BenchmarkKimi K2.6Kimi K3Leader
AA Intelligence Index27.043.6Kimi K3
Epoch Capabilities Index150.9157.9Kimi K3
Agentic workLike-for-likePublic boards, shown as published, kept out of the index213
Agentic work benchmarks for Kimi K2.6 and Kimi K3
BenchmarkKimi K2.6Kimi K3Leader
AA Agentic Index20.550.0Kimi K3
Design Arena: agents building web apps12681295Kimi K3
Design Arena: agents building full-stack apps11551315Kimi K3
Design Arena: agents building mobile apps11561269Kimi K3
Design Arena: agents building Android apps11611253Kimi K3
Design Arena: agents building Godot games11201223Kimi K3
Design Arena: agents building games11411252Kimi K3
Design Arena: HTML slides11351206Kimi K3
Design Arena: PowerPoint slides11031247Kimi K3
Design Arena: spreadsheets12761070Kimi K2.6
LiveBench researchNot measured79.0%Only Kimi K3 measured
LiveBench browserNot measured16.0%Only Kimi K3 measured
LiveBench long agenticNot measured42.8%Only Kimi K3 measured
Finance Agent (Vals AI)44.9%53.1%Kimi K3
Tax Agent Bench (Vals AI)12.4%33.4%Kimi K3
Legal Research Bench (Vals AI)15.9%46.1%Kimi K3
Harvey's Legal Agent Benchmark (Vals AI)Not measured12.9%Only Kimi K3 measured
Excel Modeling Benchmark (Vals AI)57.9%66.7%Kimi K3
BioMysteryBench (Vals AI)Not measured72.6%Only Kimi K3 measured
Terminal-Bench-Science (Vals AI)Not measured1.4%Only Kimi K3 measured
Time Horizon Index (Vals AI)Not measured10.5%Only Kimi K3 measured
τ²-Bench Telecom95.9%Not measuredOnly Kimi K2.6 measured
APEX-AgentsNot measured47.8%Only Kimi K3 measured
OSWorld 2.04.6%Not measuredOnly Kimi K2.6 measured
PostTrainBenchNot measured32.0%Only Kimi K3 measured
Vending-Bench 2$6,205$5,165Kimi K2.6
Neither model has a result on the other 15 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index010
Reasoning benchmarks for Kimi K2.6 and Kimi K3
BenchmarkKimi K2.6Kimi K3Leader
GPQA Diamond89.1%92.9%Kimi K3
LiveBench reasoningNot measured90.7%Only Kimi K3 measured
MMMU Pro (Vals AI)86.3%88.2%Kimi K3
GPQA Diamond (Artificial Analysis run)91.1%93.5%Kimi K3
Humanity's Last Exam (Artificial Analysis run)37.5%46.9%Kimi K3
GPQA Diamond (Epoch AI run)90.8%93.1%Kimi K3
ARC-AGINot measured94.5%Only Kimi K3 measured
ARC-AGI-2Not measured60.4%Only Kimi K3 measured
SimpleBenchNot measured60.7%Only Kimi K3 measured
Chess puzzles20.0%37.0%Kimi K3
Mystery game puzzles18.0%26.0%Kimi K3
EBR-bench2.4%Not measuredOnly Kimi K2.6 measured
LMArena Text: hard prompts14701495Kimi K3
LMArena Text: expert14911520Kimi K3
LMArena Text: hard prompts in English14711496Kimi K3
LMArena Vision: diagrams1298Not measuredOnly Kimi K2.6 measured
LMArena Vision: homework1310Not measuredOnly Kimi K2.6 measured
Neither model has a result on the other 5 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index06
Math benchmarks for Kimi K2.6 and Kimi K3
BenchmarkKimi K2.6Kimi K3Leader
OTIS Mock AIME96.1%97.2%Kimi K3
FrontierMath, tiers 1 to 357.2%72.2%Kimi K3
FrontierMath, tier 425.6%39.0%Kimi K3
ProofBench16.0%87.0%Kimi K3
LMArena Text: math14741494Kimi K3
LMArena Text: mathematical work14801505Kimi K3
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index011
Knowledge benchmarks for Kimi K2.6 and Kimi K3
BenchmarkKimi K2.6Kimi K3Leader
MMLU Pro87.6%88.0%Kimi K3
LegalBench (Vals AI)84.7%86.2%Kimi K3
CaseLaw (Vals AI)61.2%Not measuredOnly Kimi K2.6 measured
CorpFin (Vals AI)66.7%71.6%Kimi K3
TaxEval (Vals AI)74.7%75.7%Kimi K3
MortgageTax (Vals AI)65.8%66.3%Kimi K3
MedCode (Vals AI)40.1%49.4%Kimi K3
SimpleQA Verified34.9%50.6%Kimi K3
LMArena Text: science14681508Kimi K3
LMArena Text: medicine and healthcare14571489Kimi K3
LMArena Text: legal and government14621503Kimi K3
LMArena Text: business and finance14451470Kimi K3
Neither model has a result on the other 6 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index07
Web design benchmarks for Kimi K2.6 and Kimi K3
BenchmarkKimi K2.6Kimi K3Leader
Design Arena: websites12681334Kimi K3
Design Arena: UI components12631341Kimi K3
Design Arena: game dev12501358Kimi K3
Design Arena: data viz12491351Kimi K3
Design Arena: 3D12671387Kimi K3
Design Arena: all code categories12711360Kimi K3
Design Arena: SVG11811302Kimi K3
Design Arena: ASCII art1165Not measuredOnly Kimi K2.6 measured
LiveBench designNot measured77.2%Only Kimi K3 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index025
General benchmarks for Kimi K2.6 and Kimi K3
BenchmarkKimi K2.6Kimi K3Leader
Vals IndexNot measured53.0%Only Kimi K3 measured
LiveBench data analysisNot measured78.7%Only Kimi K3 measured
Public Benefits Bench (Vals AI)56.6%68.2%Kimi K3
SAGE (Vals AI)50.2%52.8%Kimi K3
MedScribe (Vals AI)78.2%88.0%Kimi K3
Vals Multimodal Index56.4%73.4%Kimi K3
IFBench76.0%Not measuredOnly Kimi K2.6 measured
Long context reasoning (AA-LCR)81.0%88.7%Kimi K3
DTBench90.9%91.2%Kimi K3
LMCA37.3%52.7%Kimi K3
Surface Evolver BenchNot measured95.0%Only Kimi K3 measured
Furniture assembly21.7%34.2%Kimi K3
ExploitBench18.4%Not measuredOnly Kimi K2.6 measured
LMArena Text14551476Kimi K3
LMArena Text: instruction following14511485Kimi K3
LMArena Text: creative writing14341457Kimi K3
LMArena Vision1282Not measuredOnly Kimi K2.6 measured
LMArena Text: English14601481Kimi K3
LMArena Text: non-English14461466Kimi K3
LMArena Text: Chinese15201529Kimi K3
LMArena Text: German14511490Kimi K3
LMArena Text: Japanese14431494Kimi K3
LMArena Text: Korean14271460Kimi K3
LMArena Text: French14711492Kimi K3
LMArena Text: Spanish14641472Kimi K3
LMArena Text: Russian14461485Kimi K3
LMArena Text: Polish14381458Kimi K3
LMArena Text: multi-turn14531489Kimi K3
LMArena Text: longer queries14681494Kimi K3
LMArena Text: writing and language14391471Kimi K3
LMArena Text: entertainment and media14281453Kimi K3
LMArena Vision: English1287Not measuredOnly Kimi K2.6 measured
LMArena Vision: Chinese1346Not measuredOnly Kimi K2.6 measured
LMArena Vision: reading text1291Not measuredOnly Kimi K2.6 measured
LMArena Vision: humor1280Not measuredOnly Kimi K2.6 measured
LMArena Vision: creative writing1273Not measuredOnly Kimi K2.6 measured
LMArena Document1451Not measuredOnly Kimi K2.6 measured
Neither model has a result on the other 9 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Kimi K2.6 and Kimi K3
BenchmarkKimi K2.6Kimi 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

Kimi K2.6$0.0029Fits in one request
Kimi K3$0.011Fits in one request

Kimi K2.6 costs less

Repository review

50K input + 3K output tokens

Kimi K2.6$0.059Fits in one request
Kimi K3$0.195Fits in one request

Kimi K2.6 costs less

Agent loop

200K input + 10K output tokens

Kimi K2.6$0.230Fits in one request
Kimi K3$0.750Fits in one request

Kimi K2.6 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 Kimi K2.6 and Kimi K3
SpecificationMoonshotAI Kimi K2.6MoonshotAI Kimi K3
MakerMoonshot AIMoonshot AI
WeightsOpen weightsOpen weights
ReleasedApril 20, 2026July 16, 2026
Context window262K1.05M
Input $/Mtok$0.95$3.00
Output $/Mtok$4.00$15.00
Coding index52.162.3
Independent measurements behind it36
Cost per solved task$0.35$0.81
SourceOpenRouter listingOpenRouter listing

Questions

Is Kimi K2.6 better than Kimi K3 for coding?

Pick Kimi K3 for coding: it scores 62.3 on the Vibe Coding Index against 52.1 for Kimi K2.6. Pick Kimi K2.6 for output tokens 3.8x cheaper ($4.00 vs $15.00 per million) and a lower cost per solved task ($0.35 against $0.81).

Which is cheaper, Kimi K2.6 or Kimi K3?

Kimi K2.6 costs $0.95 per million input tokens and $4.00 per million output tokens; Kimi K3 costs $3.00 and $15.00. Per solved SWE-bench Verified task, Kimi K2.6 costs $0.35 and Kimi K3 $0.81.

Which has the bigger context window, Kimi K2.6 or Kimi K3?

Kimi K3, with 1.05M tokens against 262K.

How many benchmarks is the Kimi K2.6 vs Kimi K3 comparison based on?

92 benchmarks have a result for both models as of October 22, 2026. Kimi K2.6's coding index rests on 3 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/kimi-k2-6-vs-kimi-k3" title="Kimi K2.6 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/kimi-k2-6-vs-kimi-k3">LLMBenchmark.ai, Kimi K2.6 vs Kimi K3</a></p>