Enter opens the highlighted result. Esc closes.

Gemini 4 Carbon vs Kimi K3

Gemini 4 Carbon leads Kimi K3 on coding, 86.6 to 61.4 on the Vibe Coding Index, and Gemini 4 Carbon costs 1.5x less per output token ($10.00 vs $15.00 per million). Updated November 14, 2026; 69 benchmarks have a result for both.

Replace Gemini 4 Carbon Replace Kimi K3
Gemini 4 CarbonGoogle
Coding
86.6
Overall
88.3
Output $/M
$10.00
64
Kimi K3Moonshot AI
Coding
61.4
Overall
61.2
Output $/M
$15.00
5
vs boards won of 69 shared

Gemini 4 Carbon or Kimi K3: which to pick

Pick Gemini 4 Carbon for coding: it scores 86.6 on the Vibe Coding Index against 61.4 for Kimi K3, though on 4 benchmarks against 6.

Pick Kimi K3 for work measured by LiveBench reasoning or Design Arena: websites, where it leads.

Model AGemini 4 CarbonGoogle86.6/1004 independent benchmarksNo cost per solved taskModel BKimi K3Moonshot AI61.4/1006 independent benchmarks$0.81 per solved task
Shared results
69
Gemini 4 Carbon only
4
Kimi K3 only
64
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
Gemini 4 Carbon 86.6 from 4 benchmarks, Kimi K3 61.4 from 6 benchmarks.
Widest category
Intelligence, 28.3 points apart on average over 2 shared benchmarks.
Widest single gap
Vending-Bench 2: Gemini 4 Carbon $16,484 against Kimi K3 $5,165, 67.7 points apart on a 0 to 100 scale.
Output price
Gemini 4 Carbon $10.00 against Kimi K3 $15.00 per million output tokens, 1.5x apart.
Context window
Both take 1.05M 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
CategoryGemini 4 CarbonBarsKimi K3Ties
Coding120
Intelligence20
Agentic work100
Reasoning111
Math50
Knowledge60
Web design33
General151
All shared boards645
CodingLike-for-likeSix of these boards feed the Vibe Coding Index120
Coding benchmarks for Gemini 4 Carbon and Kimi K3
BenchmarkGemini 4 CarbonKimi K3Leader
LiveBench coding91.0%53.1%Gemini 4 Carbon
LMArena WebDev17681658Gemini 4 Carbon
AA Coding IndexNot measured72.0Only Kimi K3 measured
LiveCodeBench (Vals AI)Not measured87.2%Only Kimi K3 measured
SWE-bench Verified (Vals AI)Not measured93.4%Only Kimi K3 measured
Terminal-Bench 4 (Vals AI)63.1%17.2%Gemini 4 Carbon
Vibe Code Bench (Vals AI)89.1%85.0%Gemini 4 Carbon
Code migration (Vals AI)Not measured16.1%Only Kimi K3 measured
ProgramBench (Vals AI)Not measured2.0%Only Kimi K3 measured
IOI (Vals AI)Not measured48.9%Only Kimi K3 measured
Terminal-Bench 2.1 (Vals AI)Not measured80.9%Only Kimi K3 measured
SWE-bench Multimodal (Vals AI)Not measured44.2%Only Kimi K3 measured
Vibe Code Bench 1-100 (Vals AI)Not measured18.2%Only Kimi K3 measured
CyberBench (Vals AI)Not measured75.2%Only Kimi K3 measured
SciCodeNot measured59.5%Only Kimi K3 measured
Terminal-Bench 4.0 (Artificial Analysis run)65.2%12.6%Gemini 4 Carbon
FrontierCode 2.040.6%10.9%Gemini 4 Carbon
FrontierCodeNot measured44.2%Only Kimi K3 measured
FrontierSWE61.2%25.9%Gemini 4 Carbon
WeirdMLNot measured82.6%Only Kimi K3 measured
MirrorCodeNot measured54.4%Only Kimi K3 measured
ALE-Bench2374.71524.5Gemini 4 Carbon
LMArena Text: coding15641506Gemini 4 Carbon
LMArena Text: software and IT15521497Gemini 4 Carbon
LMArena WebDev: React17921672Gemini 4 Carbon
LMArena WebDev: HTML17701640Gemini 4 Carbon
LMArena WebDev: from an imageNot measured1586Only Kimi K3 measured
SWE-Bench Pro V2Not measured97.7%Only Kimi K3 measured
GameBench 2Not measured49.8Only Kimi K3 measured
Terminal-Bench 4.0Not measured22.4%Only Kimi K3 measured
CursorBench 4.060.4%Not measuredOnly Gemini 4 Carbon measured
Neither model has a result on the other 21 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Gemini 4 Carbon and Kimi K3
BenchmarkGemini 4 CarbonKimi K3Leader
AA Intelligence Index60.017.4Gemini 4 Carbon
Epoch Capabilities Index169.8158.6Gemini 4 Carbon
Agentic workLike-for-likePublic boards, shown as published, kept out of the index100
Agentic work benchmarks for Gemini 4 Carbon and Kimi K3
BenchmarkGemini 4 CarbonKimi K3Leader
AA Agentic IndexNot measured29.3Only Kimi K3 measured
Design Arena: agents building web appsNot measured1273Only Kimi K3 measured
Design Arena: agents building full-stack appsNot measured1290Only Kimi K3 measured
Design Arena: agents building mobile apps12841253Gemini 4 Carbon
Design Arena: agents building Android appsNot measured1256Only Kimi K3 measured
Design Arena: agents building Godot gamesNot measured1200Only Kimi K3 measured
Design Arena: agents building gamesNot measured1198Only Kimi K3 measured
Design Arena: HTML slidesNot measured1151Only Kimi K3 measured
Design Arena: PowerPoint slidesNot measured1223Only Kimi K3 measured
Design Arena: spreadsheets12241068Gemini 4 Carbon
LiveBench research92.5%79.8%Gemini 4 Carbon
LiveBench browser80.0%75.2%Gemini 4 Carbon
LiveBench long agentic72.7%36.0%Gemini 4 Carbon
Finance Agent (Vals AI)Not measured53.1%Only Kimi K3 measured
Tax Agent Bench (Vals AI)51.8%33.4%Gemini 4 Carbon
Legal Research Bench (Vals AI)55.8%46.1%Gemini 4 Carbon
Harvey's Legal Agent Benchmark (Vals AI)17.5%12.9%Gemini 4 Carbon
Excel Modeling Benchmark (Vals AI)Not measured66.7%Only Kimi K3 measured
BioMysteryBench (Vals AI)Not measured72.6%Only Kimi K3 measured
Terminal-Bench-Science (Vals AI)Not measured1.4%Only Kimi K3 measured
Time Horizon Index (Vals AI)Not measured10.5%Only Kimi K3 measured
APEX-Agents84.2%47.8%Gemini 4 Carbon
PostTrainBenchNot measured31.5%Only Kimi K3 measured
Vending-Bench 2$16,484$5,165Gemini 4 Carbon
MCP AtlasNot measured82.3%Only Kimi K3 measured
Toolathlon-VerifiedNot measured76.5%Only Kimi K3 measured
τ³-BankingNot measured37.1%Only Kimi K3 measured
Neither model has a result on the other 17 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index111
Reasoning benchmarks for Gemini 4 Carbon and Kimi K3
BenchmarkGemini 4 CarbonKimi K3Leader
GPQA DiamondNot measured92.9%Only Kimi K3 measured
LiveBench reasoning90.6%90.7%Kimi K3
MMMU Pro (Vals AI)Not measured88.2%Only Kimi K3 measured
GPQA Diamond (Artificial Analysis run)Not measured93.5%Only Kimi K3 measured
Humanity's Last Exam (Artificial Analysis run)Not measured46.9%Only Kimi K3 measured
HLE-Diamond (Artificial Analysis run)69.1%26.9%Gemini 4 Carbon
CritPt v2, hard subset (Artificial Analysis run)69.0%40.5%Gemini 4 Carbon
GPQA Diamond (Epoch AI run)96.5%93.1%Gemini 4 Carbon
ARC-AGI98.5%94.5%Gemini 4 Carbon
ARC-AGI-294.2%60.4%Gemini 4 Carbon
SimpleBench92.6%60.7%Gemini 4 Carbon
Chess puzzles78.0%37.0%Gemini 4 Carbon
Mystery game puzzles73.0%26.0%Gemini 4 Carbon
EBR-bench87.6%Not measuredOnly Gemini 4 Carbon measured
LMArena Text: hard prompts15511494Gemini 4 Carbon
LMArena Text: expert15821515Gemini 4 Carbon
LMArena Text: hard prompts in English15451493Gemini 4 Carbon
Humanity's Last Exam (Scale run)Not measured22.2%Only Kimi K3 measured
Neither model has a result on the other 8 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index50
Math benchmarks for Gemini 4 Carbon and Kimi K3
BenchmarkGemini 4 CarbonKimi K3Leader
OTIS Mock AIME100.0%97.2%Gemini 4 Carbon
FrontierMath, tiers 1 to 390.5%72.2%Gemini 4 Carbon
FrontierMath, tier 495.1%39.0%Gemini 4 Carbon
ProofBenchNot measured87.0%Only Kimi K3 measured
LMArena Text: mathNot measured1496Only Kimi K3 measured
LMArena Text: mathematical work15381507Gemini 4 Carbon
MathArena85.4%42.9%Gemini 4 Carbon
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index60
Knowledge benchmarks for Gemini 4 Carbon and Kimi K3
BenchmarkGemini 4 CarbonKimi K3Leader
MMLU ProNot measured88.0%Only Kimi K3 measured
LegalBench (Vals AI)Not measured86.2%Only Kimi K3 measured
CorpFin (Vals AI)Not measured71.6%Only Kimi K3 measured
TaxEval (Vals AI)Not measured75.7%Only Kimi K3 measured
MortgageTax (Vals AI)Not measured66.3%Only Kimi K3 measured
MedCode (Vals AI)Not measured49.4%Only Kimi K3 measured
AA-Omniscience53.5-1.7Gemini 4 Carbon
SimpleQA Verified77.7%50.6%Gemini 4 Carbon
LMArena Text: science15351502Gemini 4 Carbon
LMArena Text: medicine and healthcare15251488Gemini 4 Carbon
LMArena Text: legal and government15421498Gemini 4 Carbon
LMArena Text: business and finance15201469Gemini 4 Carbon
FACTS Benchmark Suite78.4%Not measuredOnly Gemini 4 Carbon measured
PRBench Legal53.3%Not measuredOnly Gemini 4 Carbon measured
Neither model has a result on the other 7 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index33
Web design benchmarks for Gemini 4 Carbon and Kimi K3
BenchmarkGemini 4 CarbonKimi K3Leader
Design Arena: websites13191336Kimi K3
Design Arena: UI components13571340Gemini 4 Carbon
Design Arena: game dev13261353Kimi K3
Design Arena: data vizNot measured1350Only Kimi K3 measured
Design Arena: 3D14341364Gemini 4 Carbon
Design Arena: all code categories13481351Kimi K3
Design Arena: SVGNot measured1296Only Kimi K3 measured
LiveBench design80.2%72.5%Gemini 4 Carbon
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index151
General benchmarks for Gemini 4 Carbon and Kimi K3
BenchmarkGemini 4 CarbonKimi K3Leader
Vals Index70.5%53.0%Gemini 4 Carbon
LiveBench data analysis86.5%78.7%Gemini 4 Carbon
Public Benefits Bench (Vals AI)Not measured68.2%Only Kimi K3 measured
SAGE (Vals AI)Not measured86.8%Only Kimi K3 measured
MedScribe (Vals AI)Not measured88.0%Only Kimi K3 measured
Vals Multimodal IndexNot measured73.4%Only Kimi K3 measured
Long context reasoning (AA-LCR)79.3%88.7%Kimi K3
GDPval-AA17331542Gemini 4 Carbon
DTBenchNot measured91.2%Only Kimi K3 measured
LMCANot measured52.7%Only Kimi K3 measured
Surface Evolver BenchNot measured95.0%Only Kimi K3 measured
Furniture assembly77.5%34.2%Gemini 4 Carbon
LMArena Text15381475Gemini 4 Carbon
LMArena Text: instruction following15391484Gemini 4 Carbon
LMArena Text: creative writing15331453Gemini 4 Carbon
LMArena Text: English15301479Gemini 4 Carbon
LMArena Text: non-English15341465Gemini 4 Carbon
LMArena Text: Chinese16131527Gemini 4 Carbon
LMArena Text: GermanNot measured1483Only Kimi K3 measured
LMArena Text: JapaneseNot measured1495Only Kimi K3 measured
LMArena Text: KoreanNot measured1444Only Kimi K3 measured
LMArena Text: FrenchNot measured1482Only Kimi K3 measured
LMArena Text: SpanishNot measured1477Only Kimi K3 measured
LMArena Text: Russian15601481Gemini 4 Carbon
LMArena Text: PolishNot measured1472Only Kimi K3 measured
LMArena Text: multi-turn15691489Gemini 4 Carbon
LMArena Text: longer queries15491490Gemini 4 Carbon
LMArena Text: writing and language15421468Gemini 4 Carbon
LMArena Text: entertainment and media15271450Gemini 4 Carbon
EQ-Bench 4Not measured1319Only Kimi K3 measured
Neither model has a result on the other 19 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Gemini 4 Carbon and Kimi K3
BenchmarkGemini 4 CarbonKimi 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

Gemini 4 Carbon$0.0070Fits in one request
Kimi K3$0.011Fits in one request

Gemini 4 Carbon costs less

Repository review

50K input + 3K output tokens

Gemini 4 Carbon$0.130Fits in one request
Kimi K3$0.195Fits in one request

Gemini 4 Carbon costs less

Agent loop

200K input + 10K output tokens

Gemini 4 Carbon$0.500Fits in one request
Kimi K3$0.750Fits in one request

Gemini 4 Carbon 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 Gemini 4 Carbon and Kimi K3
SpecificationGoogle Gemini 4 CarbonMoonshotAI Kimi K3
MakerGoogleMoonshot AI
WeightsNo public weights listedOpen weights
ReleasedNovember 10, 2026July 16, 2026
Context window1.05M1.05M
Input $/Mtok$2.00$3.00
Output $/Mtok$10.00$15.00
Coding index86.661.4
Independent measurements behind it46
Cost per solved taskNot computable$0.81
SourceOpenRouter listingOpenRouter listing

Questions

Is Gemini 4 Carbon better than Kimi K3 for coding?

Pick Gemini 4 Carbon for coding: it scores 86.6 on the Vibe Coding Index against 61.4 for Kimi K3, though on 4 benchmarks against 6. Pick Kimi K3 for work measured by LiveBench reasoning or Design Arena: websites, where it leads.

Which is cheaper, Gemini 4 Carbon or Kimi K3?

Gemini 4 Carbon costs $2.00 per million input tokens and $10.00 per million output tokens; Kimi K3 costs $3.00 and $15.00. Cost per solved task is known for Kimi K3 only.

Which has the bigger context window, Gemini 4 Carbon or Kimi K3?

Neither: both take 1.05M tokens.

How many benchmarks is the Gemini 4 Carbon vs Kimi K3 comparison based on?

69 benchmarks have a result for both models as of November 14, 2026. Gemini 4 Carbon's coding index rests on 4 independent benchmarks and Kimi K3's on 6. 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/gemini-4-carbon-vs-kimi-k3" title="Gemini 4 Carbon 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/gemini-4-carbon-vs-kimi-k3">LLMBenchmark.ai, Gemini 4 Carbon vs Kimi K3</a></p>