LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

GPT-5.6 Luna vs Kimi K3

Updated October 21, 2026. 101 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace GPT-5.6 Luna Replace Kimi K3
GPT-5.6 LunaOpenAI
Coding
61.5
Overall
37.3
Output $/M
$1.20
15
Kimi K3Moonshot AI
Coding
61.8
Overall
43.6
Output $/M
$11.936
86
vs boards won of 101 shared

Decision reading

GPT-5.6 Luna and Kimi K3 are level for coding, 61.5 against 61.8 on the Vibe Coding Index, so price and context decide.

Pick GPT-5.6 Luna when output tokens cost 9.9x less, or when a solved task costs $0.04 against $0.81. Pick Kimi K3 when what matters is LMArena WebDev or AA Coding Index, where it leads.

Model AGPT-5.6 LunaOpenAI61.5/1002 independent benchmarks$0.04 per solved taskModel BKimi K3Moonshot AI61.8/1002 independent benchmarks$0.81 per solved task
Shared results
101
GPT-5.6 Luna only
15
Kimi K3 only
16
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
GPT-5.6 Luna 61.5 from 2 benchmarks, Kimi K3 61.8 from 2 benchmarks.
Widest category
Web design, 15.5 points apart on average over 8 shared benchmarks.
Widest single gap
LiveBench browser: GPT-5.6 Luna 68.0% against Kimi K3 16.0%, 52.0 points apart on a 0 to 100 scale.
Output price
GPT-5.6 Luna $1.20 against Kimi K3 $11.936 per million output tokens, 9.9x apart.
Context window
GPT-5.6 Luna takes 1.05M 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
CategoryGPT-5.6 LunaBarsKimi K3Ties
Coding516
Intelligence02
Agentic work39
Reasoning113
Math33
Knowledge29
Web design08
General126
All shared boards1586
CodingLike-for-likeFour of these boards feed the Vibe Coding Index516
Coding benchmarks for GPT-5.6 Luna and Kimi K3
BenchmarkGPT-5.6 LunaKimi K3Leader
LiveBench coding73.6%59.3%GPT-5.6 Luna
LMArena WebDev15181655Kimi K3
AA Coding Index71.476.2Kimi K3
LiveCodeBench (Vals AI)Not measured87.2%Only Kimi K3 measured
SWE-bench Verified (Vals AI)93.0%93.4%Kimi K3
Terminal-Bench 4 (Vals AI)11.6%17.2%Kimi K3
Vibe Code Bench (Vals AI)77.1%85.0%Kimi K3
Code migration (Vals AI)44.5%16.1%GPT-5.6 Luna
ProgramBench (Vals AI)0.0%2.0%Kimi K3
IOI (Vals AI)61.8%48.9%GPT-5.6 Luna
Terminal-Bench 2.1 (Vals AI)79.0%80.9%Kimi K3
SWE-bench Multimodal (Vals AI)Not measured44.2%Only Kimi K3 measured
Vibe Code Bench 1-100 (Vals AI)22.6%18.2%GPT-5.6 Luna
SkillsBench (Vals AI)60.5%Not measuredOnly GPT-5.6 Luna measured
CyberBench (Vals AI)73.6%75.2%Kimi K3
SciCode53.6%59.5%Kimi K3
FrontierCode39.8%44.2%Kimi K3
FrontierSWENot measured25.9%Only Kimi K3 measured
WeirdML60.9%82.6%Kimi K3
ALE-Bench1667.41524.5GPT-5.6 Luna
LMArena Text: coding14661508Kimi K3
LMArena Text: software and IT14631499Kimi K3
LMArena WebDev: React15081670Kimi K3
LMArena WebDev: HTML15201644Kimi K3
LMArena WebDev: from an image15171579Kimi K3
Neither model has a result on the other 13 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for GPT-5.6 Luna and Kimi K3
BenchmarkGPT-5.6 LunaKimi K3Leader
AA Intelligence Index37.343.6Kimi K3
Epoch Capabilities Index156.5157.9Kimi K3
Agentic workLike-for-likePublic boards, shown as published, kept out of the index39
Agentic work benchmarks for GPT-5.6 Luna and Kimi K3
BenchmarkGPT-5.6 LunaKimi K3Leader
AA Agentic Index42.150.0Kimi K3
Design Arena: agents building web appsNot measured1296Only Kimi K3 measured
Design Arena: agents building full-stack appsNot measured1315Only Kimi K3 measured
Design Arena: agents building mobile appsNot measured1267Only Kimi K3 measured
Design Arena: agents building Android appsNot measured1256Only 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 research51.0%79.0%Kimi K3
LiveBench browser68.0%16.0%GPT-5.6 Luna
LiveBench long agentic30.4%42.8%Kimi K3
SREBench (Vals AI)1.5%Not measuredOnly GPT-5.6 Luna measured
Finance Agent (Vals AI)55.0%53.1%GPT-5.6 Luna
Tax Agent Bench (Vals AI)24.8%33.4%Kimi K3
Legal Research Bench (Vals AI)36.5%46.1%Kimi K3
Harvey's Legal Agent Benchmark (Vals AI)Not measured12.9%Only Kimi K3 measured
Excel Modeling Benchmark (Vals AI)67.1%66.7%GPT-5.6 Luna
BioMysteryBench (Vals AI)61.5%72.6%Kimi K3
Terminal-Bench-Science (Vals AI)0.0%1.4%Kimi K3
Vals Science Index25.3%Not measuredOnly GPT-5.6 Luna measured
Time Horizon Index (Vals AI)Not measured10.5%Only Kimi K3 measured
APEX-Agents43.0%47.8%Kimi K3
BALROG45.6%Not measuredOnly GPT-5.6 Luna measured
PostTrainBenchNot measured32.0%Only Kimi K3 measured
Vending-Bench 2$4,095$5,165Kimi K3
Neither model has a result on the other 12 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index113
Reasoning benchmarks for GPT-5.6 Luna and Kimi K3
BenchmarkGPT-5.6 LunaKimi K3Leader
GPQA Diamond91.7%92.9%Kimi K3
LiveBench reasoning85.6%90.7%Kimi K3
MysteryMechanism (Vals AI)14.4%Not measuredOnly GPT-5.6 Luna measured
MMMU Pro (Vals AI)85.0%88.2%Kimi K3
GPQA Diamond (Artificial Analysis run)91.1%93.5%Kimi K3
Humanity's Last Exam (Artificial Analysis run)39.5%46.9%Kimi K3
GPQA Diamond (Epoch AI run)91.6%93.1%Kimi K3
ARC-AGI88.0%94.5%Kimi K3
ARC-AGI-259.5%60.4%Kimi K3
SimpleBench46.8%60.7%Kimi K3
Chess puzzles43.0%37.0%GPT-5.6 Luna
Mystery game puzzles21.0%26.0%Kimi K3
LMArena Text: hard prompts14511495Kimi K3
LMArena Text: expert14781520Kimi K3
LMArena Text: hard prompts in English14601496Kimi K3
LMArena Vision: diagrams1273Not measuredOnly GPT-5.6 Luna measured
LMArena Vision: homework1282Not measuredOnly GPT-5.6 Luna measured
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index33
Math benchmarks for GPT-5.6 Luna and Kimi K3
BenchmarkGPT-5.6 LunaKimi K3Leader
OTIS Mock AIME98.3%97.2%GPT-5.6 Luna
FrontierMath, tiers 1 to 382.1%72.2%GPT-5.6 Luna
FrontierMath, tier 461.0%39.0%GPT-5.6 Luna
ProofBench60.0%87.0%Kimi K3
LMArena Text: math14591494Kimi K3
LMArena Text: mathematical work14701505Kimi K3
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index29
Knowledge benchmarks for GPT-5.6 Luna and Kimi K3
BenchmarkGPT-5.6 LunaKimi K3Leader
MMLU Pro86.0%88.0%Kimi K3
LegalBench (Vals AI)84.0%86.2%Kimi K3
CorpFin (Vals AI)64.2%71.6%Kimi K3
TaxEval (Vals AI)76.2%75.7%GPT-5.6 Luna
MortgageTax (Vals AI)67.3%66.3%GPT-5.6 Luna
MedCode (Vals AI)42.4%49.4%Kimi K3
SimpleQA Verified41.0%50.6%Kimi K3
LMArena Text: science14331508Kimi K3
LMArena Text: medicine and healthcare13991489Kimi K3
LMArena Text: legal and government14461503Kimi K3
LMArena Text: business and finance14391470Kimi K3
Neither model has a result on the other 6 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index08
Web design benchmarks for GPT-5.6 Luna and Kimi K3
BenchmarkGPT-5.6 LunaKimi K3Leader
Design Arena: websites12561336Kimi K3
Design Arena: UI components12501345Kimi K3
Design Arena: game dev11961361Kimi K3
Design Arena: data viz12121350Kimi K3
Design Arena: 3D11211388Kimi K3
Design Arena: all code categories12311361Kimi K3
Design Arena: SVG11931301Kimi K3
Design Arena: ASCII art1144Not measuredOnly GPT-5.6 Luna measured
LiveBench design76.0%77.2%Kimi K3
GeneralLike-for-likePublic boards, shown as published, kept out of the index126
General benchmarks for GPT-5.6 Luna and Kimi K3
BenchmarkGPT-5.6 LunaKimi K3Leader
Vals IndexNot measured53.0%Only Kimi K3 measured
LiveBench data analysis78.0%78.7%Kimi K3
Public Benefits Bench (Vals AI)61.2%68.2%Kimi K3
SAGE (Vals AI)44.2%52.8%Kimi K3
MedScribe (Vals AI)84.4%88.0%Kimi K3
Vals Multimodal Index69.1%73.4%Kimi K3
Long context reasoning (AA-LCR)83.7%88.7%Kimi K3
DTBench89.1%91.2%Kimi K3
LMCA48.5%52.7%Kimi K3
Surface Evolver Bench61.9%95.0%Kimi K3
Furniture assembly42.5%34.2%GPT-5.6 Luna
LMArena Text14311476Kimi K3
LMArena Text: instruction following14371485Kimi K3
LMArena Text: creative writing13961457Kimi K3
LMArena Vision1257Not measuredOnly GPT-5.6 Luna measured
LMArena Text: English14411481Kimi K3
LMArena Text: non-English14171466Kimi K3
LMArena Text: Chinese14701529Kimi K3
LMArena Text: German14551490Kimi K3
LMArena Text: Japanese14151494Kimi K3
LMArena Text: Korean14151460Kimi K3
LMArena Text: French14581492Kimi K3
LMArena Text: Spanish14461472Kimi K3
LMArena Text: Russian14271485Kimi K3
LMArena Text: Polish14051458Kimi K3
LMArena Text: multi-turn14341489Kimi K3
LMArena Text: longer queries14361494Kimi K3
LMArena Text: writing and language14191471Kimi K3
LMArena Text: entertainment and media14021453Kimi K3
LMArena Vision: English1263Not measuredOnly GPT-5.6 Luna measured
LMArena Vision: Chinese1294Not measuredOnly GPT-5.6 Luna measured
LMArena Vision: reading text1264Not measuredOnly GPT-5.6 Luna measured
LMArena Vision: humor1190Not measuredOnly GPT-5.6 Luna measured
LMArena Vision: creative writing1257Not measuredOnly GPT-5.6 Luna measured
LMArena Document1457Not measuredOnly GPT-5.6 Luna 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 GPT-5.6 Luna and Kimi K3
BenchmarkGPT-5.6 LunaKimi 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

GPT-5.6 Luna$0.0008Fits in one request
Kimi K3$0.0062Fits in one request

GPT-5.6 Luna costs less

Repository review

50K input + 3K output tokens

GPT-5.6 Luna$0.014Fits in one request
Kimi K3$0.048Fits in one request

GPT-5.6 Luna costs less

Agent loop

200K input + 10K output tokens

GPT-5.6 Luna$0.052Fits in one request
Kimi K3$0.167Fits in one request

GPT-5.6 Luna 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 GPT-5.6 Luna and Kimi K3
SpecificationOpenAI GPT-5.6 LunaMoonshotAI Kimi K3
MakerOpenAIMoonshot AI
WeightsNo public weights listedOpen weights
ReleasedJuly 9, 2026July 16, 2026
Context window1.05M1.05M
Input $/Mtok$0.20$0.239
Output $/Mtok$1.20$11.936
Coding index61.561.8
Independent measurements behind it22
Cost per solved task$0.04$0.81
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5.6 Luna better than Kimi K3 for coding?

GPT-5.6 Luna and Kimi K3 are level for coding, 61.5 against 61.8 on the Vibe Coding Index, so price and context decide. Pick GPT-5.6 Luna when output tokens cost 9.9x less, or when a solved task costs $0.04 against $0.81. Pick Kimi K3 when what matters is LMArena WebDev or AA Coding Index, where it leads.

Which is cheaper, GPT-5.6 Luna or Kimi K3?

GPT-5.6 Luna costs $0.20 per million input tokens and $1.20 per million output tokens; Kimi K3 costs $0.239 and $11.936. Per solved SWE-bench Verified task, GPT-5.6 Luna costs $0.04 and Kimi K3 $0.81.

Which has the bigger context window, GPT-5.6 Luna or Kimi K3?

GPT-5.6 Luna, with 1.05M tokens against 1.05M.

How many benchmarks is the GPT-5.6 Luna vs Kimi K3 comparison based on?

101 benchmarks have a result for both models as of October 21, 2026. GPT-5.6 Luna's coding index rests on 2 independent benchmarks and Kimi K3'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/gpt-5-6-luna-vs-kimi-k3" title="GPT-5.6 Luna 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/gpt-5-6-luna-vs-kimi-k3">LLMBenchmark.ai, GPT-5.6 Luna vs Kimi K3</a></p>