LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

GPT-6 Luna vs Inkling Small

Updated October 21, 2026. 70 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 Inkling Small
GPT-6 LunaOpenAI
Coding
71.4
Overall
38.1
Output $/M
$0.50
40
T Inkling SmallThinking Machines
Coding
29.9
Overall
25.7
Output $/M
$1.20
29
vs boards won of 70 shared, 1 tied

Decision reading

Pick GPT-6 Luna for coding: it scores 71.4 on the Vibe Coding Index against 29.9 for Inkling Small.

Pick Inkling Small 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 BInkling SmallThinking Machines29.9/1001 independent benchmark$0.16 per solved task
Shared results
70
GPT-6 Luna only
32
Inkling Small only
14
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
GPT-6 Luna 71.4 from 2 benchmarks, Inkling Small 29.9 from 1 benchmark.
Widest category
Math, 24.9 points apart on average over 6 shared benchmarks.
Widest single gap
Vibe Code Bench (Vals AI): GPT-6 Luna 81.7% against Inkling Small 19.1%, 62.6 points apart on a 0 to 100 scale.
Output price
GPT-6 Luna $0.50 against Inkling Small $1.20 per million output tokens, 2.4x apart.
Context window
GPT-6 Luna takes 1.05M tokens, Inkling Small 524K.

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 LunaBarsInkling SmallTies
Coding1121 tie
Intelligence20
Agentic work50
Reasoning73
Math42
Knowledge24
General918
All shared boards40291 tie
CodingLike-for-likeFour of these boards feed the Vibe Coding Index112
Coding benchmarks for GPT-6 Luna and Inkling Small
BenchmarkGPT-6 LunaInkling SmallLeader
LiveBench coding84.5%Not measuredOnly GPT-6 Luna measured
LMArena WebDev15781409GPT-6 Luna
AA Coding IndexNot measured52.9Only Inkling Small measured
LiveCodeBench (Vals AI)Not measured85.9%Only Inkling Small measured
SWE-bench Verified (Vals AI)Not measured82.2%Only Inkling Small measured
Terminal-Bench 4 (Vals AI)13.6%0.0%GPT-6 Luna
Vibe Code Bench (Vals AI)81.7%19.1%GPT-6 Luna
Code migration (Vals AI)42.5%13.7%GPT-6 Luna
ProgramBench (Vals AI)0.5%0.5%Tie
IOI (Vals AI)55.6%9.3%GPT-6 Luna
Terminal-Bench 2.1 (Vals AI)73.0%55.1%GPT-6 Luna
SWE-bench Multimodal (Vals AI)37.5%Not measuredOnly GPT-6 Luna measured
SkillsBench (Vals AI)Not measured33.6%Only Inkling Small measured
CyberBench (Vals AI)76.3%64.1%GPT-6 Luna
SciCode54.6%49.7%GPT-6 Luna
FrontierCode42.4%Not measuredOnly GPT-6 Luna measured
ALE-Bench1576.9Not measuredOnly GPT-6 Luna measured
LMArena Text: coding14371450Inkling Small
LMArena Text: software and IT14231447Inkling Small
LMArena WebDev: React15731399GPT-6 Luna
LMArena WebDev: HTML15951398GPT-6 Luna
LMArena WebDev: from an image15711410GPT-6 Luna
Neither model has a result on the other 16 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GPT-6 Luna and Inkling Small
BenchmarkGPT-6 LunaInkling SmallLeader
AA Intelligence Index38.125.7GPT-6 Luna
Epoch Capabilities Index156.8150.3GPT-6 Luna
Agentic workLike-for-likePublic boards, shown as published, kept out of the index50
Agentic work benchmarks for GPT-6 Luna and Inkling Small
BenchmarkGPT-6 LunaInkling SmallLeader
AA Agentic IndexNot measured23.5Only Inkling Small measured
Design Arena: agents building web apps1189Not measuredOnly GPT-6 Luna measured
Design Arena: agents building full-stack apps1158Not measuredOnly GPT-6 Luna measured
Design Arena: agents building mobile apps1133Not measuredOnly GPT-6 Luna 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%41.3%GPT-6 Luna
Tax Agent Bench (Vals AI)19.9%15.4%GPT-6 Luna
Legal Research Bench (Vals AI)30.3%25.5%GPT-6 Luna
Harvey's Legal Agent Benchmark (Vals AI)4.2%2.9%GPT-6 Luna
Excel Modeling Benchmark (Vals AI)68.5%31.8%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
APEX-Agents44.3%Not measuredOnly GPT-6 Luna measured
SpireClimb0.0%Not measuredOnly GPT-6 Luna measured
Neither model has a result on the other 21 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index73
Reasoning benchmarks for GPT-6 Luna and Inkling Small
BenchmarkGPT-6 LunaInkling SmallLeader
GPQA DiamondNot measured83.6%Only Inkling Small measured
LiveBench reasoning81.8%Not measuredOnly GPT-6 Luna measured
MysteryMechanism (Vals AI)19.4%Not measuredOnly GPT-6 Luna measured
GPQA Diamond (Artificial Analysis run)Not measured89.5%Only Inkling Small measured
Humanity's Last Exam (Artificial Analysis run)38.5%33.3%GPT-6 Luna
GPQA Diamond (Epoch AI run)90.5%88.5%GPT-6 Luna
ARC-AGI86.7%84.0%GPT-6 Luna
ARC-AGI-259.3%40.1%GPT-6 Luna
Chess puzzles36.0%Not measuredOnly GPT-6 Luna measured
Mystery game puzzles7.0%6.0%GPT-6 Luna
LMArena Text: hard prompts14101423Inkling Small
LMArena Text: expert14421441GPT-6 Luna
LMArena Text: hard prompts in English14101429Inkling Small
LMArena Vision: diagrams12451251Inkling Small
LMArena Vision: homework12611257GPT-6 Luna
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index42
Math benchmarks for GPT-6 Luna and Inkling Small
BenchmarkGPT-6 LunaInkling SmallLeader
OTIS Mock AIME98.9%90.0%GPT-6 Luna
FrontierMath, tiers 1 to 378.9%46.3%GPT-6 Luna
FrontierMath, tier 456.1%17.1%GPT-6 Luna
ProofBench64.0%6.0%GPT-6 Luna
LMArena Text: math14091459Inkling Small
LMArena Text: mathematical work14271441Inkling Small
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index24
Knowledge benchmarks for GPT-6 Luna and Inkling Small
BenchmarkGPT-6 LunaInkling SmallLeader
MMLU ProNot measured85.6%Only Inkling Small measured
LegalBench (Vals AI)Not measured83.0%Only Inkling Small measured
CorpFin (Vals AI)Not measured69.6%Only Inkling Small measured
TaxEval (Vals AI)Not measured75.5%Only Inkling Small measured
MortgageTax (Vals AI)Not measured62.3%Only Inkling Small measured
MedCode (Vals AI)44.7%37.9%GPT-6 Luna
SimpleQA Verified41.4%19.1%GPT-6 Luna
LMArena Text: science13861428Inkling Small
LMArena Text: medicine and healthcare13521413Inkling Small
LMArena Text: legal and government14001415Inkling Small
LMArena Text: business and finance13881413Inkling Small
Neither model has a result on the other 6 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
Web design benchmarks for GPT-6 Luna and Inkling Small
BenchmarkGPT-6 LunaInkling SmallLeader
Design Arena: websites1272Not measuredOnly GPT-6 Luna measured
Design Arena: UI components1297Not measuredOnly GPT-6 Luna measured
Design Arena: game dev1259Not measuredOnly GPT-6 Luna measured
Design Arena: data viz1215Not measuredOnly GPT-6 Luna measured
Design Arena: 3D1217Not measuredOnly GPT-6 Luna measured
Design Arena: all code categories1271Not measuredOnly GPT-6 Luna measured
Design Arena: SVG1253Not measuredOnly GPT-6 Luna measured
Design Arena: ASCII art1156Not measuredOnly GPT-6 Luna measured
LiveBench design82.5%Not measuredOnly GPT-6 Luna measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index918
General benchmarks for GPT-6 Luna and Inkling Small
BenchmarkGPT-6 LunaInkling SmallLeader
Vals Index52.3%27.6%GPT-6 Luna
LiveBench data analysis73.4%Not measuredOnly GPT-6 Luna measured
Public Benefits Bench (Vals AI)57.6%59.4%Inkling Small
SAGE (Vals AI)48.1%32.7%GPT-6 Luna
MedScribe (Vals AI)83.7%84.1%Inkling Small
Vals Multimodal IndexNot measured50.0%Only Inkling Small measured
Long context reasoning (AA-LCR)83.3%75.7%GPT-6 Luna
DTBench90.1%Not measuredOnly GPT-6 Luna measured
LMCA44.5%Not measuredOnly GPT-6 Luna measured
Furniture assembly44.2%Not measuredOnly GPT-6 Luna measured
LMArena Text13891413Inkling Small
LMArena Text: instruction following14081399GPT-6 Luna
LMArena Text: creative writing13631334GPT-6 Luna
LMArena Vision12161237Inkling Small
LMArena Text: English13931421Inkling Small
LMArena Text: non-English13811401Inkling Small
LMArena Text: Chinese14271464Inkling Small
LMArena Text: German13761408Inkling Small
LMArena Text: Japanese13691401Inkling Small
LMArena Text: Korean13591367Inkling Small
LMArena Text: French14081435Inkling Small
LMArena Text: Spanish13911422Inkling Small
LMArena Text: Russian13921394Inkling Small
LMArena Text: Polish13831428Inkling Small
LMArena Text: multi-turn13921416Inkling Small
LMArena Text: longer queries14081402GPT-6 Luna
LMArena Text: writing and language13841357GPT-6 Luna
LMArena Text: entertainment and media13641352GPT-6 Luna
LMArena Vision: English12041243Inkling Small
LMArena Vision: Chinese12661292Inkling Small
LMArena Vision: reading text12251243Inkling Small
LMArena Vision: humorNot measured1207Only Inkling Small measured
LMArena Vision: creative writing12051205GPT-6 Luna
Neither model has a result on the other 13 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 Inkling Small
BenchmarkGPT-6 LunaInkling SmallLeader
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
Inkling Small$0.0010Fits in one request

GPT-6 Luna costs less

Repository review

50K input + 3K output tokens

GPT-6 Luna$0.0065Fits in one request
Inkling Small$0.026Fits in one request

GPT-6 Luna costs less

Agent loop

200K input + 10K output tokens

GPT-6 Luna$0.025Fits in one request
Inkling Small$0.102Fits 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 Inkling Small
SpecificationOpenAI GPT-6 LunaThinking Machines Inkling Small
MakerOpenAIThinking Machines
WeightsNo public weights listedOpen weights
ReleasedSeptember 22, 2026July 30, 2026
Context window1.05M524K
Input $/Mtok$0.10$0.45
Output $/Mtok$0.50$1.20
Coding index71.429.9
Independent measurements behind it21
Cost per solved taskNot computable$0.16
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-6 Luna better than Inkling Small for coding?

Pick GPT-6 Luna for coding: it scores 71.4 on the Vibe Coding Index against 29.9 for Inkling Small. Pick Inkling Small when what matters is LMArena Text: coding or LMArena Text: software and IT, where it leads.

Which is cheaper, GPT-6 Luna or Inkling Small?

GPT-6 Luna costs $0.10 per million input tokens and $0.50 per million output tokens; Inkling Small costs $0.45 and $1.20. Cost per solved task is known for Inkling Small only.

Which has the bigger context window, GPT-6 Luna or Inkling Small?

GPT-6 Luna, with 1.05M tokens against 524K.

How many benchmarks is the GPT-6 Luna vs Inkling Small comparison based on?

70 benchmarks have a result for both models as of October 21, 2026. GPT-6 Luna's coding index rests on 2 independent benchmarks and Inkling Small's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next