LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Kimi K2.6 vs Qwen3.5-27B

Updated October 21, 2026. 51 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Kimi K2.6 Replace Qwen3.5-27B
Kimi K2.6Moonshot AI
Coding
44.2
Overall
27.0
Output $/M
$3.41
51
Qwen3.5-27BQwen
Coding
22.6
Overall
22.9
Output $/M
$2.16
0
vs boards won of 51 shared

Decision reading

Pick Kimi K2.6 for coding: it scores 44.2 on the Vibe Coding Index against 22.6 for Qwen3.5-27B.

Pick Qwen3.5-27B when output tokens cost 1.6x less.

Model AKimi K2.6Moonshot AI44.2/1001 independent benchmark$0.35 per solved taskModel BQwen3.5-27BQwen22.6/1001 independent benchmarkNo cost per solved task
Shared results
51
Kimi K2.6 only
58
Qwen3.5-27B only
2
Like-for-like categories
7 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
Kimi K2.6 44.2 from 1 benchmark, Qwen3.5-27B 22.6 from 1 benchmark.
Widest category
Agentic work, 19.8 points apart on average over 2 shared benchmarks.
Widest single gap
Vending-Bench 2: Kimi K2.6 $6,205 against Qwen3.5-27B $202, 37.5 points apart on a 0 to 100 scale.
Output price
Kimi K2.6 $3.41 against Qwen3.5-27B $2.16 per million output tokens, 1.6x apart.
Context window
Both take 262K 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
CategoryKimi K2.6BarsQwen3.5-27BTies
Coding80
Intelligence10
Agentic work20
Reasoning70
Math20
Knowledge40
General270
All shared boards510
CodingLike-for-likeFour of these boards feed the Vibe Coding Index80
Coding benchmarks for Kimi K2.6 and Qwen3.5-27B
BenchmarkKimi K2.6Qwen3.5-27BLeader
LMArena WebDev15091358Kimi K2.6
AA Coding Index61.8Not measuredOnly Kimi K2.6 measured
LiveCodeBench (Vals AI)86.8%Not measuredOnly Kimi K2.6 measured
SWE-bench Verified (Vals AI)76.2%Not measuredOnly Kimi K2.6 measured
Vibe Code Bench (Vals AI)37.9%Not measuredOnly Kimi K2.6 measured
Code migration (Vals AI)27.8%Not measuredOnly Kimi K2.6 measured
ProgramBench (Vals AI)0.0%Not measuredOnly Kimi K2.6 measured
Terminal-Bench 2.1 (Vals AI)53.6%Not measuredOnly Kimi K2.6 measured
SciCode51.5%Not measuredOnly Kimi K2.6 measured
Terminal-Bench Hard (Artificial Analysis run)43.9%32.6%Kimi K2.6
WeirdML55.9%39.5%Kimi K2.6
ALE-Bench1092.7349.4Kimi K2.6
LMArena Text: coding14891427Kimi K2.6
LMArena Text: software and IT14841431Kimi K2.6
LMArena WebDev: React15051346Kimi K2.6
LMArena WebDev: HTML14851386Kimi K2.6
LMArena WebDev: from an image1521Not measuredOnly Kimi K2.6 measured
Neither model has a result on the other 21 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
Intelligence benchmarks for Kimi K2.6 and Qwen3.5-27B
BenchmarkKimi K2.6Qwen3.5-27BLeader
AA Intelligence Index27.022.9Kimi K2.6
Epoch Capabilities Index150.9Not measuredOnly Kimi K2.6 measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index20
Agentic work benchmarks for Kimi K2.6 and Qwen3.5-27B
BenchmarkKimi K2.6Qwen3.5-27BLeader
AA Agentic Index20.5Not measuredOnly Kimi K2.6 measured
Design Arena: agents building web apps1268Not measuredOnly Kimi K2.6 measured
Design Arena: agents building full-stack apps1155Not measuredOnly Kimi K2.6 measured
Design Arena: agents building mobile apps1156Not measuredOnly Kimi K2.6 measured
Design Arena: agents building Android apps1166Not measuredOnly Kimi K2.6 measured
Design Arena: agents building Godot games1120Not measuredOnly Kimi K2.6 measured
Design Arena: agents building games1141Not measuredOnly Kimi K2.6 measured
Design Arena: HTML slides1184Not measuredOnly Kimi K2.6 measured
Design Arena: PowerPoint slides1128Not measuredOnly Kimi K2.6 measured
Design Arena: spreadsheets1276Not measuredOnly Kimi K2.6 measured
Finance Agent (Vals AI)44.9%Not measuredOnly Kimi K2.6 measured
Tax Agent Bench (Vals AI)12.4%Not measuredOnly Kimi K2.6 measured
Legal Research Bench (Vals AI)15.9%Not measuredOnly Kimi K2.6 measured
Excel Modeling Benchmark (Vals AI)57.9%Not measuredOnly Kimi K2.6 measured
τ²-Bench Telecom95.9%93.9%Kimi K2.6
OSWorld 2.04.6%Not measuredOnly Kimi K2.6 measured
Vending-Bench 2$6,205$202Kimi K2.6
Neither model has a result on the other 22 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index70
Reasoning benchmarks for Kimi K2.6 and Qwen3.5-27B
BenchmarkKimi K2.6Qwen3.5-27BLeader
GPQA Diamond89.1%Not measuredOnly Kimi K2.6 measured
MMMU Pro (Vals AI)86.3%Not measuredOnly Kimi K2.6 measured
GPQA Diamond (Artificial Analysis run)91.1%85.8%Kimi K2.6
Humanity's Last Exam (Artificial Analysis run)37.5%23.9%Kimi K2.6
GPQA Diamond (Epoch AI run)90.8%Not measuredOnly Kimi K2.6 measured
Chess puzzles20.0%Not measuredOnly Kimi K2.6 measured
Mystery game puzzles18.0%Not measuredOnly Kimi K2.6 measured
EBR-bench2.4%Not measuredOnly Kimi K2.6 measured
LMArena Text: hard prompts14701415Kimi K2.6
LMArena Text: expert14911427Kimi K2.6
LMArena Text: hard prompts in English14711429Kimi K2.6
LMArena Vision: diagrams12981249Kimi K2.6
LMArena Vision: homework13101256Kimi K2.6
Neither model has a result on the other 8 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index20
Math benchmarks for Kimi K2.6 and Qwen3.5-27B
BenchmarkKimi K2.6Qwen3.5-27BLeader
OTIS Mock AIME96.1%Not measuredOnly Kimi K2.6 measured
FrontierMath, tiers 1 to 357.2%Not measuredOnly Kimi K2.6 measured
FrontierMath, tier 425.6%Not measuredOnly Kimi K2.6 measured
ProofBench16.0%Not measuredOnly Kimi K2.6 measured
LMArena Text: math14741429Kimi K2.6
LMArena Text: mathematical work14801429Kimi K2.6
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index40
Knowledge benchmarks for Kimi K2.6 and Qwen3.5-27B
BenchmarkKimi K2.6Qwen3.5-27BLeader
MMLU Pro87.6%Not measuredOnly Kimi K2.6 measured
LegalBench (Vals AI)84.7%Not measuredOnly Kimi K2.6 measured
CaseLaw (Vals AI)61.2%Not measuredOnly Kimi K2.6 measured
CorpFin (Vals AI)66.7%Not measuredOnly Kimi K2.6 measured
TaxEval (Vals AI)74.7%Not measuredOnly Kimi K2.6 measured
MortgageTax (Vals AI)65.8%Not measuredOnly Kimi K2.6 measured
MedCode (Vals AI)40.1%Not measuredOnly Kimi K2.6 measured
SimpleQA Verified34.9%Not measuredOnly Kimi K2.6 measured
LMArena Text: science14681437Kimi K2.6
LMArena Text: medicine and healthcare14571425Kimi K2.6
LMArena Text: legal and government14621405Kimi K2.6
LMArena Text: business and finance14451399Kimi K2.6
Neither model has a result on the other 5 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
Web design benchmarks for Kimi K2.6 and Qwen3.5-27B
BenchmarkKimi K2.6Qwen3.5-27BLeader
Design Arena: websites1269Not measuredOnly Kimi K2.6 measured
Design Arena: UI components1266Not measuredOnly Kimi K2.6 measured
Design Arena: game dev1251Not measuredOnly Kimi K2.6 measured
Design Arena: data viz1249Not measuredOnly Kimi K2.6 measured
Design Arena: 3D1268Not measuredOnly Kimi K2.6 measured
Design Arena: all code categories1272Not measuredOnly Kimi K2.6 measured
Design Arena: SVG1181Not measuredOnly Kimi K2.6 measured
Design Arena: ASCII art1164Not measuredOnly Kimi K2.6 measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index270
General benchmarks for Kimi K2.6 and Qwen3.5-27B
BenchmarkKimi K2.6Qwen3.5-27BLeader
Public Benefits Bench (Vals AI)56.6%Not measuredOnly Kimi K2.6 measured
SAGE (Vals AI)50.2%Not measuredOnly Kimi K2.6 measured
MedScribe (Vals AI)78.2%Not measuredOnly Kimi K2.6 measured
Vals Multimodal Index56.4%Not measuredOnly Kimi K2.6 measured
IFBench76.0%75.6%Kimi K2.6
Long context reasoning (AA-LCR)81.0%77.7%Kimi K2.6
DTBench90.9%82.4%Kimi K2.6
LMCA37.3%34.0%Kimi K2.6
Furniture assembly21.7%Not measuredOnly Kimi K2.6 measured
ExploitBench18.4%Not measuredOnly Kimi K2.6 measured
LMArena Text14551409Kimi K2.6
LMArena Text: instruction following14511393Kimi K2.6
LMArena Text: creative writing14341362Kimi K2.6
LMArena Vision12821241Kimi K2.6
LMArena Text: English14601425Kimi K2.6
LMArena Text: non-English14461390Kimi K2.6
LMArena Text: Chinese15201478Kimi K2.6
LMArena Text: German14511392Kimi K2.6
LMArena Text: Japanese14431346Kimi K2.6
LMArena Text: Korean14271359Kimi K2.6
LMArena Text: French14711410Kimi K2.6
LMArena Text: Spanish14641407Kimi K2.6
LMArena Text: Russian14461390Kimi K2.6
LMArena Text: Polish14381375Kimi K2.6
LMArena Text: multi-turn14531410Kimi K2.6
LMArena Text: longer queries14681413Kimi K2.6
LMArena Text: writing and language14391380Kimi K2.6
LMArena Text: entertainment and media14281363Kimi K2.6
LMArena Vision: English12871246Kimi K2.6
LMArena Vision: Chinese13461301Kimi K2.6
LMArena Vision: reading text12911246Kimi K2.6
LMArena Vision: humor12801227Kimi K2.6
LMArena Vision: captioningNot measured1170Only Qwen3.5-27B measured
LMArena Vision: recognizing thingsNot measured1210Only Qwen3.5-27B measured
LMArena Vision: creative writing12731224Kimi K2.6
LMArena Document1451Not measuredOnly Kimi K2.6 measured
Neither model has a result on the other 10 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 Qwen3.5-27B
BenchmarkKimi K2.6Qwen3.5-27BLeader
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$0.0024Fits in one request
Qwen3.5-27B$0.0014Fits in one request

Qwen3.5-27B costs less

Repository review

50K input + 3K output tokens

Kimi K2.6$0.043Fits in one request
Qwen3.5-27B$0.020Fits in one request

Qwen3.5-27B costs less

Agent loop

200K input + 10K output tokens

Kimi K2.6$0.164Fits in one request
Qwen3.5-27B$0.076Fits in one request

Qwen3.5-27B 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 Qwen3.5-27B
SpecificationMoonshotAI Kimi K2.6Qwen3.5-27B
MakerMoonshot AIQwen
WeightsOpen weightsOpen weights
ReleasedApril 20, 2026February 25, 2026
Context window262K262K
Input $/Mtok$0.65$0.27
Output $/Mtok$3.41$2.16
Coding index44.222.6
Independent measurements behind it11
Cost per solved task$0.35Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Kimi K2.6 better than Qwen3.5-27B for coding?

Pick Kimi K2.6 for coding: it scores 44.2 on the Vibe Coding Index against 22.6 for Qwen3.5-27B. Pick Qwen3.5-27B when output tokens cost 1.6x less.

Which is cheaper, Kimi K2.6 or Qwen3.5-27B?

Kimi K2.6 costs $0.65 per million input tokens and $3.41 per million output tokens; Qwen3.5-27B costs $0.27 and $2.16. Cost per solved task is known for Kimi K2.6 only.

Which has the bigger context window, Kimi K2.6 or Qwen3.5-27B?

Neither: both take 262K tokens.

How many benchmarks is the Kimi K2.6 vs Qwen3.5-27B comparison based on?

51 benchmarks have a result for both models as of October 21, 2026. Kimi K2.6's coding index rests on 1 independent benchmark and Qwen3.5-27B's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next