LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

GPT-6 Luna vs Kimi K2.5

Updated October 21, 2026. 67 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace GPT-6 Luna Replace Kimi K2.5
GPT-6 LunaOpenAI
Coding
71.4
Overall
38.1
Output $/M
$0.50
28
Kimi K2.5Moonshot AI
Coding
58.5
Overall
23.5
Output $/M
$2.50
39
vs boards won of 67 shared

Decision reading

Pick GPT-6 Luna for coding: it scores 71.4 on the Vibe Coding Index against 58.5 for Kimi K2.5.

Pick Kimi K2.5 when what matters is LMArena Text: coding or LMArena Text: software and IT, where it leads.

Model AGPT-6 LunaOpenAI71.4/1002 independent benchmarksNo cost per solved taskModel BKimi K2.5Moonshot AI58.5/1002 independent benchmarks$0.24 per solved task
Shared results
67
GPT-6 Luna only
35
Kimi K2.5 only
31
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-6 Luna 71.4 from 2 benchmarks, Kimi K2.5 58.5 from 2 benchmarks.
Widest category
Coding, 26.6 points apart on average over 10 shared benchmarks.
Widest single gap
Vibe Code Bench (Vals AI): GPT-6 Luna 81.7% against Kimi K2.5 17.5%, 64.1 points apart on a 0 to 100 scale.
Output price
GPT-6 Luna $0.50 against Kimi K2.5 $2.50 per million output tokens, 5.0x apart.
Context window
GPT-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-6 LunaBarsKimi K2.5Ties
Coding82
Intelligence20
Agentic work51
Reasoning35
Math02
Knowledge24
Web design62
General223
All shared boards2839
CodingLike-for-likeFour of these boards feed the Vibe Coding Index82
Coding benchmarks for GPT-6 Luna and Kimi K2.5
BenchmarkGPT-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 coding84.5%Not measuredOnly GPT-6 Luna measured
LMArena WebDev15781437GPT-6 Luna
AA Coding IndexNot measured46.8Only Kimi K2.5 measured
LiveCodeBench (Vals AI)Not measured83.9%Only Kimi K2.5 measured
SWE-bench Verified (Vals AI)Not measured70.0%Only Kimi K2.5 measured
Terminal-Bench 4 (Vals AI)13.6%Not measuredOnly GPT-6 Luna measured
Vibe Code Bench (Vals AI)81.7%17.5%GPT-6 Luna
Code migration (Vals AI)42.5%7.0%GPT-6 Luna
ProgramBench (Vals AI)0.5%Not measuredOnly GPT-6 Luna measured
IOI (Vals AI)55.6%Not measuredOnly GPT-6 Luna measured
Terminal-Bench 2.1 (Vals AI)73.0%42.0%GPT-6 Luna
SWE-bench Multimodal (Vals AI)37.5%Not measuredOnly GPT-6 Luna measured
CyberBench (Vals AI)76.3%Not measuredOnly GPT-6 Luna measured
SciCode54.6%Not measuredOnly GPT-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
FrontierCode42.4%Not measuredOnly GPT-6 Luna measured
WeirdMLNot measured45.6%Only Kimi K2.5 measured
ALE-Bench1576.9821.6GPT-6 Luna
LMArena Text: coding14371483Kimi K2.5
LMArena Text: software and IT14231471Kimi K2.5
LMArena WebDev: React15731429GPT-6 Luna
LMArena WebDev: HTML15951426GPT-6 Luna
LMArena WebDev: from an image15711439GPT-6 Luna
Neither model has a result on the other 12 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GPT-6 Luna and Kimi K2.5
BenchmarkGPT-6 LunaKimi K2.5Leader
AA Intelligence Index38.123.5GPT-6 Luna
Epoch Capabilities Index156.8148.1GPT-6 Luna
Agentic workLike-for-likePublic boards, shown as published, kept out of the index51
Agentic work benchmarks for GPT-6 Luna and Kimi K2.5
BenchmarkGPT-6 LunaKimi K2.5Leader
Design Arena: agents building web apps11891136GPT-6 Luna
Design Arena: agents building full-stack apps11581114GPT-6 Luna
Design Arena: agents building mobile apps11331147Kimi K2.5
Design Arena: agents building Android appsNot measured1124Only Kimi K2.5 measured
Design Arena: agents building Godot gamesNot measured1219Only Kimi K2.5 measured
LiveBench research43.5%Not measuredOnly GPT-6 Luna measured
LiveBench browser72.0%Not measuredOnly GPT-6 Luna measured
LiveBench long agentic48.1%Not measuredOnly GPT-6 Luna measured
SREBench (Vals AI)2.7%Not measuredOnly GPT-6 Luna measured
Finance Agent (Vals AI)49.9%35.8%GPT-6 Luna
Tax Agent Bench (Vals AI)19.9%Not measuredOnly GPT-6 Luna measured
Legal Research Bench (Vals AI)30.3%15.9%GPT-6 Luna
Harvey's Legal Agent Benchmark (Vals AI)4.2%Not measuredOnly GPT-6 Luna measured
Excel Modeling Benchmark (Vals AI)68.5%28.5%GPT-6 Luna
BioMysteryBench (Vals AI)61.5%Not measuredOnly GPT-6 Luna measured
Terminal-Bench-Science (Vals AI)4.3%Not measuredOnly GPT-6 Luna measured
Vals Science Index28.4%Not measuredOnly GPT-6 Luna measured
τ²-Bench TelecomNot measured95.9%Only Kimi K2.5 measured
APEX-Agents44.3%Not measuredOnly GPT-6 Luna measured
OSWorldNot measured63.3%Only Kimi K2.5 measured
SpireClimb0.0%Not measuredOnly GPT-6 Luna measured
Vending-Bench 2Not measured$1,198Only Kimi K2.5 measured
Neither model has a result on the other 17 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index35
Reasoning benchmarks for GPT-6 Luna and Kimi K2.5
BenchmarkGPT-6 LunaKimi K2.5Leader
GPQA DiamondNot measured84.1%Only Kimi K2.5 measured
LiveBench reasoning81.8%Not measuredOnly GPT-6 Luna measured
MysteryMechanism (Vals AI)19.4%Not measuredOnly GPT-6 Luna measured
MMMU Pro (Vals AI)Not measured84.3%Only Kimi K2.5 measured
GPQA Diamond (Artificial Analysis run)Not measured87.9%Only Kimi K2.5 measured
Humanity's Last Exam (Artificial Analysis run)38.5%30.7%GPT-6 Luna
GPQA Diamond (Epoch AI run)90.5%Not measuredOnly GPT-6 Luna measured
ARC-AGI86.7%65.3%GPT-6 Luna
ARC-AGI-259.3%11.8%GPT-6 Luna
SimpleBenchNot measured46.8%Only Kimi K2.5 measured
Chess puzzles36.0%Not measuredOnly GPT-6 Luna measured
Mystery game puzzles7.0%Not measuredOnly GPT-6 Luna measured
LMArena Text: hard prompts14101453Kimi K2.5
LMArena Text: expert14421466Kimi K2.5
LMArena Text: hard prompts in English14101460Kimi K2.5
LMArena Vision: diagrams12451282Kimi K2.5
LMArena Vision: homework12611297Kimi 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-6 Luna and Kimi K2.5
BenchmarkGPT-6 LunaKimi K2.5Leader
AIMENot measured95.6%Only Kimi K2.5 measured
OTIS Mock AIME98.9%Not measuredOnly GPT-6 Luna measured
FrontierMath, tiers 1 to 378.9%Not measuredOnly GPT-6 Luna measured
FrontierMath, tier 456.1%Not measuredOnly GPT-6 Luna measured
ProofBench64.0%Not measuredOnly GPT-6 Luna measured
LMArena Text: math14091470Kimi K2.5
LMArena Text: mathematical work14271473Kimi K2.5
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index24
Knowledge benchmarks for GPT-6 Luna and Kimi K2.5
BenchmarkGPT-6 LunaKimi K2.5Leader
MMLU ProNot measured85.9%Only Kimi K2.5 measured
CaseLaw (Vals AI)Not measured58.7%Only Kimi K2.5 measured
CorpFin (Vals AI)Not measured68.3%Only Kimi K2.5 measured
TaxEval (Vals AI)Not measured74.2%Only Kimi K2.5 measured
MortgageTax (Vals AI)Not measured66.5%Only Kimi K2.5 measured
MedQA (Vals AI)Not measured94.4%Only Kimi K2.5 measured
MedCode (Vals AI)44.7%39.3%GPT-6 Luna
SimpleQA Verified41.4%34.3%GPT-6 Luna
LMArena Text: science13861459Kimi K2.5
LMArena Text: medicine and healthcare13521448Kimi K2.5
LMArena Text: legal and government14001441Kimi K2.5
LMArena Text: business and finance13881427Kimi K2.5
Neither model has a result on the other 5 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index62
Web design benchmarks for GPT-6 Luna and Kimi K2.5
BenchmarkGPT-6 LunaKimi K2.5Leader
Design Arena: websites12721249GPT-6 Luna
Design Arena: UI components12971236GPT-6 Luna
Design Arena: game dev12591207GPT-6 Luna
Design Arena: data viz12151219Kimi K2.5
Design Arena: 3D12171205GPT-6 Luna
Design Arena: all code categories12711236GPT-6 Luna
Design Arena: SVG12531149GPT-6 Luna
Design Arena: ASCII art11561165Kimi K2.5
LiveBench design82.5%Not measuredOnly GPT-6 Luna measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index223
General benchmarks for GPT-6 Luna and Kimi K2.5
BenchmarkGPT-6 LunaKimi K2.5Leader
Vals Index52.3%Not measuredOnly GPT-6 Luna measured
LiveBench data analysis73.4%Not measuredOnly GPT-6 Luna measured
Public Benefits Bench (Vals AI)57.6%Not measuredOnly GPT-6 Luna measured
SAGE (Vals AI)48.1%49.9%Kimi K2.5
MedScribe (Vals AI)83.7%76.4%GPT-6 Luna
IFBenchNot measured70.2%Only Kimi K2.5 measured
Long context reasoning (AA-LCR)83.3%78.0%GPT-6 Luna
DTBench90.1%Not measuredOnly GPT-6 Luna measured
LMCA44.5%Not measuredOnly GPT-6 Luna measured
Fiction.LiveBenchNot measured86.1%Only Kimi K2.5 measured
CL-benchNot measured19.3%Only Kimi K2.5 measured
Furniture assembly44.2%Not measuredOnly GPT-6 Luna measured
CL-bench LifeNot measured13.2%Only Kimi K2.5 measured
LMArena Text13891445Kimi K2.5
LMArena Text: instruction following14081431Kimi K2.5
LMArena Text: creative writing13631422Kimi K2.5
LMArena Vision12161269Kimi K2.5
LMArena Text: English13931453Kimi K2.5
LMArena Text: non-English13811433Kimi K2.5
LMArena Text: Chinese14271494Kimi K2.5
LMArena Text: German13761441Kimi K2.5
LMArena Text: Japanese13691423Kimi K2.5
LMArena Text: Korean13591412Kimi K2.5
LMArena Text: French14081454Kimi K2.5
LMArena Text: Spanish13911450Kimi K2.5
LMArena Text: Russian13921435Kimi K2.5
LMArena Text: Polish13831445Kimi K2.5
LMArena Text: multi-turn13921444Kimi K2.5
LMArena Text: longer queries14081445Kimi K2.5
LMArena Text: writing and language13841425Kimi K2.5
LMArena Text: entertainment and media13641417Kimi K2.5
LMArena Vision: English12041274Kimi K2.5
LMArena Vision: Chinese12661328Kimi K2.5
LMArena Vision: reading text12251276Kimi K2.5
LMArena Vision: humorNot measured1284Only Kimi K2.5 measured
LMArena Vision: captioningNot measured1257Only Kimi K2.5 measured
LMArena Vision: recognizing thingsNot measured1285Only Kimi K2.5 measured
LMArena Vision: creative writing12051269Kimi K2.5
Neither model has a result on the other 8 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-6 Luna and Kimi K2.5
BenchmarkGPT-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-6 Luna$0.0003Fits in one request
Kimi K2.5$0.0018Fits in one request

GPT-6 Luna costs less

Repository review

50K input + 3K output tokens

GPT-6 Luna$0.0065Fits in one request
Kimi K2.5$0.033Fits in one request

GPT-6 Luna costs less

Agent loop

200K input + 10K output tokens

GPT-6 Luna$0.025Fits in one request
Kimi K2.5$0.125Fits in one request

GPT-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-6 Luna and Kimi K2.5
SpecificationOpenAI GPT-6 LunaMoonshotAI Kimi K2.5
MakerOpenAIMoonshot AI
WeightsNo public weights listedOpen weights
ReleasedSeptember 22, 2026January 27, 2026
Context window1.05M262K
Input $/Mtok$0.10$0.50
Output $/Mtok$0.50$2.50
Coding index71.458.5
Independent measurements behind it22
Cost per solved taskNot computable$0.24
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-6 Luna better than Kimi K2.5 for coding?

Pick GPT-6 Luna for coding: it scores 71.4 on the Vibe Coding Index against 58.5 for Kimi K2.5. Pick Kimi K2.5 when what matters is LMArena Text: coding or LMArena Text: software and IT, where it leads.

Which is cheaper, GPT-6 Luna or Kimi K2.5?

GPT-6 Luna costs $0.10 per million input tokens and $0.50 per million output tokens; Kimi K2.5 costs $0.50 and $2.50. Cost per solved task is known for Kimi K2.5 only.

Which has the bigger context window, GPT-6 Luna or Kimi K2.5?

GPT-6 Luna, with 1.05M tokens against 262K.

How many benchmarks is the GPT-6 Luna vs Kimi K2.5 comparison based on?

67 benchmarks have a result for both models as of October 21, 2026. GPT-6 Luna's coding index rests on 2 independent benchmarks and Kimi K2.5'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-6-luna-vs-kimi-k2-5" title="GPT-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-6-luna-vs-kimi-k2-5">LLMBenchmark.ai, GPT-6 Luna vs Kimi K2.5</a></p>