Enter opens the highlighted result. Esc closes.

GPT-5.6 Luna vs Kimi K2.5

GPT-5.6 Luna leads Kimi K2.5 on coding, 58.0 to 38.6 on the Vibe Coding Index, and GPT-5.6 Luna costs 2.4x less per output token ($1.20 vs $2.85 per million). Updated October 22, 2026; 78 benchmarks have a result for both.

Replace GPT-5.6 Luna Replace Kimi K2.5
GPT-5.6 LunaOpenAI
Coding
58.0
Overall
61.6
Output $/M
$1.20
41
Kimi K2.5Moonshot AI
Coding
38.6
Overall
44.8
Output $/M
$2.85
37
vs boards won of 78 shared

GPT-5.6 Luna or Kimi K2.5: which to pick

Pick GPT-5.6 Luna for coding: it scores 58.0 on the Vibe Coding Index against 38.6 for Kimi K2.5.

Pick Kimi K2.5 for work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.

Model AGPT-5.6 LunaOpenAI58.0/1006 independent benchmarks$0.04 per solved taskModel BKimi K2.5Moonshot AI38.6/1003 independent benchmarks$0.24 per solved task
Shared results
78
GPT-5.6 Luna only
41
Kimi K2.5 only
21
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
GPT-5.6 Luna 58.0 from 6 benchmarks, Kimi K2.5 38.6 from 3 benchmarks.
Widest category
Agentic work, 24.2 points apart on average over 4 shared benchmarks.
Widest single gap
Vibe Code Bench (Vals AI): GPT-5.6 Luna 77.1% against Kimi K2.5 17.5%, 59.5 points apart on a 0 to 100 scale.
Output price
GPT-5.6 Luna $1.20 against Kimi K2.5 $2.85 per million output tokens, 2.4x apart.
Context window
GPT-5.6 Luna takes 1.05M tokens, Kimi K2.5 262K.

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 K2.5Ties
Coding112
Intelligence20
Agentic work40
Reasoning85
Math02
Knowledge73
Web design35
General620
All shared boards4137
CodingLike-for-likeSix of these boards feed the Vibe Coding Index112
Coding benchmarks for GPT-5.6 Luna and Kimi K2.5
BenchmarkGPT-5.6 LunaKimi K2.5Leader
SWE-bench VerifiedNot measured70.8%Only Kimi K2.5 measured
SWE-bench MultilingualNot measured67.3%Only Kimi K2.5 measured
LiveBench coding73.6%Not measuredOnly GPT-5.6 Luna measured
LMArena WebDev15181437GPT-5.6 Luna
AA Coding Index71.446.8GPT-5.6 Luna
LiveCodeBench (Vals AI)Not measured83.9%Only Kimi K2.5 measured
SWE-bench Verified (Vals AI)93.0%70.0%GPT-5.6 Luna
Terminal-Bench 4 (Vals AI)11.6%Not measuredOnly GPT-5.6 Luna measured
Vibe Code Bench (Vals AI)77.1%17.5%GPT-5.6 Luna
Code migration (Vals AI)44.5%7.0%GPT-5.6 Luna
ProgramBench (Vals AI)0.0%Not measuredOnly GPT-5.6 Luna measured
IOI (Vals AI)61.8%Not measuredOnly GPT-5.6 Luna measured
Terminal-Bench 2.1 (Vals AI)79.0%42.0%GPT-5.6 Luna
Vibe Code Bench 1-100 (Vals AI)22.6%Not measuredOnly GPT-5.6 Luna measured
SkillsBench (Vals AI)60.5%Not measuredOnly GPT-5.6 Luna measured
CyberBench (Vals AI)73.6%Not measuredOnly GPT-5.6 Luna measured
SciCode53.6%Not measuredOnly GPT-5.6 Luna measured
Terminal-Bench Hard (Artificial Analysis run)Not measured34.8%Only Kimi K2.5 measured
Terminal-BenchNot measured43.2%Only Kimi K2.5 measured
FrontierCode39.8%Not measuredOnly GPT-5.6 Luna measured
WeirdML60.9%45.6%GPT-5.6 Luna
ALE-Bench1667.4821.6GPT-5.6 Luna
LMArena Text: coding14661483Kimi K2.5
LMArena Text: software and IT14631471Kimi K2.5
LMArena WebDev: React15081429GPT-5.6 Luna
LMArena WebDev: HTML15201426GPT-5.6 Luna
LMArena WebDev: from an image15171439GPT-5.6 Luna
SciCode6.2%Not measuredOnly GPT-5.6 Luna measured
Neither model has a result on the other 11 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GPT-5.6 Luna and Kimi K2.5
BenchmarkGPT-5.6 LunaKimi K2.5Leader
AA Intelligence Index37.323.5GPT-5.6 Luna
Epoch Capabilities Index156.5148.1GPT-5.6 Luna
Agentic workLike-for-likePublic boards, shown as published, kept out of the index40
Agentic work benchmarks for GPT-5.6 Luna and Kimi K2.5
BenchmarkGPT-5.6 LunaKimi K2.5Leader
AA Agentic Index42.1Not measuredOnly GPT-5.6 Luna measured
Design Arena: agents building web appsNot measured1136Only Kimi K2.5 measured
Design Arena: agents building full-stack appsNot measured1114Only Kimi K2.5 measured
Design Arena: agents building mobile appsNot measured1147Only Kimi K2.5 measured
Design Arena: agents building Android appsNot measured1118Only Kimi K2.5 measured
Design Arena: agents building Godot gamesNot measured1219Only Kimi K2.5 measured
LiveBench research51.0%Not measuredOnly GPT-5.6 Luna measured
LiveBench browser68.0%Not measuredOnly GPT-5.6 Luna measured
LiveBench long agentic30.4%Not measuredOnly GPT-5.6 Luna measured
SREBench (Vals AI)1.5%Not measuredOnly GPT-5.6 Luna measured
Finance Agent (Vals AI)55.0%35.8%GPT-5.6 Luna
Tax Agent Bench (Vals AI)24.8%Not measuredOnly GPT-5.6 Luna measured
Legal Research Bench (Vals AI)36.5%15.9%GPT-5.6 Luna
Excel Modeling Benchmark (Vals AI)67.1%28.5%GPT-5.6 Luna
BioMysteryBench (Vals AI)61.5%Not measuredOnly GPT-5.6 Luna measured
Terminal-Bench-Science (Vals AI)0.0%Not measuredOnly GPT-5.6 Luna measured
Vals Science Index25.3%Not measuredOnly GPT-5.6 Luna measured
τ²-Bench TelecomNot measured95.9%Only Kimi K2.5 measured
APEX-Agents43.0%Not measuredOnly GPT-5.6 Luna measured
BALROG45.6%Not measuredOnly GPT-5.6 Luna measured
OSWorldNot measured63.3%Only Kimi K2.5 measured
Vending-Bench 2$4,095$1,198GPT-5.6 Luna
Neither model has a result on the other 19 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index85
Reasoning benchmarks for GPT-5.6 Luna and Kimi K2.5
BenchmarkGPT-5.6 LunaKimi K2.5Leader
GPQA Diamond91.7%84.1%GPT-5.6 Luna
LiveBench reasoning85.6%Not measuredOnly GPT-5.6 Luna measured
MysteryMechanism (Vals AI)14.4%Not measuredOnly GPT-5.6 Luna measured
MMMU Pro (Vals AI)85.0%84.3%GPT-5.6 Luna
GPQA Diamond (Artificial Analysis run)91.1%87.9%GPT-5.6 Luna
Humanity's Last Exam (Artificial Analysis run)39.5%30.7%GPT-5.6 Luna
GPQA Diamond (Epoch AI run)91.6%Not measuredOnly GPT-5.6 Luna measured
ARC-AGI88.0%65.3%GPT-5.6 Luna
ARC-AGI-259.5%11.8%GPT-5.6 Luna
SimpleBench46.8%46.8%GPT-5.6 Luna
Chess puzzles43.0%Not measuredOnly GPT-5.6 Luna measured
Mystery game puzzles21.0%Not measuredOnly GPT-5.6 Luna measured
LMArena Text: hard prompts14511453Kimi K2.5
LMArena Text: expert14781466GPT-5.6 Luna
LMArena Text: hard prompts in English14601460Kimi K2.5
LMArena Vision: diagrams12731282Kimi K2.5
LMArena Vision: homework12821297Kimi K2.5
Kagi LLM Benchmark49.1%78.5%Kimi K2.5
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index02
Math benchmarks for GPT-5.6 Luna and Kimi K2.5
BenchmarkGPT-5.6 LunaKimi K2.5Leader
AIMENot measured95.6%Only Kimi K2.5 measured
OTIS Mock AIME98.3%Not measuredOnly GPT-5.6 Luna measured
FrontierMath, tiers 1 to 382.1%Not measuredOnly GPT-5.6 Luna measured
FrontierMath, tier 461.0%Not measuredOnly GPT-5.6 Luna measured
ProofBench60.0%Not measuredOnly GPT-5.6 Luna measured
LMArena Text: math14591470Kimi K2.5
LMArena Text: mathematical work14701473Kimi K2.5
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index73
Knowledge benchmarks for GPT-5.6 Luna and Kimi K2.5
BenchmarkGPT-5.6 LunaKimi K2.5Leader
MMLU Pro86.0%85.9%GPT-5.6 Luna
LegalBench (Vals AI)84.0%Not measuredOnly GPT-5.6 Luna measured
CaseLaw (Vals AI)Not measured58.7%Only Kimi K2.5 measured
CorpFin (Vals AI)64.2%68.3%Kimi K2.5
TaxEval (Vals AI)76.2%74.2%GPT-5.6 Luna
MortgageTax (Vals AI)67.3%66.5%GPT-5.6 Luna
MedQA (Vals AI)Not measured94.4%Only Kimi K2.5 measured
MedCode (Vals AI)42.4%39.3%GPT-5.6 Luna
SimpleQA Verified41.0%34.3%GPT-5.6 Luna
LMArena Text: science14331459Kimi K2.5
LMArena Text: medicine and healthcare13991448Kimi K2.5
LMArena Text: legal and government14461441GPT-5.6 Luna
LMArena Text: business and finance14391427GPT-5.6 Luna
FACTS Benchmark Suite35.0%Not measuredOnly GPT-5.6 Luna measured
Neither model has a result on the other 4 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index35
Web design benchmarks for GPT-5.6 Luna and Kimi K2.5
BenchmarkGPT-5.6 LunaKimi K2.5Leader
Design Arena: websites12551248GPT-5.6 Luna
Design Arena: UI components12481234GPT-5.6 Luna
Design Arena: game dev11941206Kimi K2.5
Design Arena: data viz12111219Kimi K2.5
Design Arena: 3D11211204Kimi K2.5
Design Arena: all code categories12311236Kimi K2.5
Design Arena: SVG11941149GPT-5.6 Luna
Design Arena: ASCII art11421166Kimi K2.5
LiveBench design76.0%Not measuredOnly GPT-5.6 Luna measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index620
General benchmarks for GPT-5.6 Luna and Kimi K2.5
BenchmarkGPT-5.6 LunaKimi K2.5Leader
LiveBench data analysis78.0%Not measuredOnly GPT-5.6 Luna measured
Public Benefits Bench (Vals AI)61.2%Not measuredOnly GPT-5.6 Luna measured
SAGE (Vals AI)44.2%49.9%Kimi K2.5
MedScribe (Vals AI)84.4%76.4%GPT-5.6 Luna
Vals Multimodal Index69.1%Not measuredOnly GPT-5.6 Luna measured
IFBenchNot measured70.2%Only Kimi K2.5 measured
Long context reasoning (AA-LCR)83.7%78.0%GPT-5.6 Luna
DTBench89.1%Not measuredOnly GPT-5.6 Luna measured
LMCA48.5%Not measuredOnly GPT-5.6 Luna measured
Fiction.LiveBenchNot measured86.1%Only Kimi K2.5 measured
CL-benchNot measured19.3%Only Kimi K2.5 measured
Surface Evolver Bench61.9%Not measuredOnly GPT-5.6 Luna measured
Furniture assembly42.5%Not measuredOnly GPT-5.6 Luna measured
CL-bench LifeNot measured13.2%Only Kimi K2.5 measured
LMArena Text14311445Kimi K2.5
LMArena Text: instruction following14371431GPT-5.6 Luna
LMArena Text: creative writing13961422Kimi K2.5
LMArena Vision12571269Kimi K2.5
LMArena Text: English14411453Kimi K2.5
LMArena Text: non-English14171433Kimi K2.5
LMArena Text: Chinese14701494Kimi K2.5
LMArena Text: German14551441GPT-5.6 Luna
LMArena Text: Japanese14151423Kimi K2.5
LMArena Text: Korean14151412GPT-5.6 Luna
LMArena Text: French14581454GPT-5.6 Luna
LMArena Text: Spanish14461450Kimi K2.5
LMArena Text: Russian14271435Kimi K2.5
LMArena Text: Polish14051445Kimi K2.5
LMArena Text: multi-turn14341444Kimi K2.5
LMArena Text: longer queries14361445Kimi K2.5
LMArena Text: writing and language14191425Kimi K2.5
LMArena Text: entertainment and media14021417Kimi K2.5
LMArena Vision: English12631274Kimi K2.5
LMArena Vision: Chinese12941328Kimi K2.5
LMArena Vision: reading text12641276Kimi K2.5
LMArena Vision: humor11901284Kimi K2.5
LMArena Vision: captioningNot measured1257Only Kimi K2.5 measured
LMArena Vision: recognizing thingsNot measured1285Only Kimi K2.5 measured
LMArena Vision: creative writing12571269Kimi K2.5
LMArena Document1457Not measuredOnly GPT-5.6 Luna measured
Neither model has a result on the other 6 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 K2.5
BenchmarkGPT-5.6 LunaKimi K2.5Leader
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 K2.5$0.0020Fits 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 K2.5$0.035Fits 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 K2.5$0.135Fits 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 K2.5
SpecificationOpenAI GPT-5.6 LunaMoonshotAI Kimi K2.5
MakerOpenAIMoonshot AI
WeightsNo public weights listedOpen weights
ReleasedJuly 9, 2026January 27, 2026
Context window1.05M262K
Input $/Mtok$0.20$0.532
Output $/Mtok$1.20$2.85
Coding index58.038.6
Independent measurements behind it63
Cost per solved task$0.04$0.24
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5.6 Luna better than Kimi K2.5 for coding?

Pick GPT-5.6 Luna for coding: it scores 58.0 on the Vibe Coding Index against 38.6 for Kimi K2.5. Pick Kimi K2.5 for work measured by LMArena Text: coding or LMArena Text: software and IT, where it leads.

Which is cheaper, GPT-5.6 Luna or Kimi K2.5?

GPT-5.6 Luna costs $0.20 per million input tokens and $1.20 per million output tokens; Kimi K2.5 costs $0.532 and $2.85. Per solved SWE-bench Verified task, GPT-5.6 Luna costs $0.04 and Kimi K2.5 $0.24.

Which has the bigger context window, GPT-5.6 Luna or Kimi K2.5?

GPT-5.6 Luna, with 1.05M tokens against 262K.

How many benchmarks is the GPT-5.6 Luna vs Kimi K2.5 comparison based on?

78 benchmarks have a result for both models as of October 22, 2026. GPT-5.6 Luna's coding index rests on 6 independent benchmarks and Kimi K2.5's on 3. 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-k2-5" title="GPT-5.6 Luna vs Kimi K2.5, 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-k2-5">LLMBenchmark.ai, GPT-5.6 Luna vs Kimi K2.5</a></p>