Enter opens the highlighted result. Esc closes.

Gemini 3.8 Flash vs Kimi K3

Kimi K3 leads Gemini 3.8 Flash on coding, 62.3 to 61.6 on the Vibe Coding Index, and Gemini 3.8 Flash costs 4.0x less per output token ($3.75 vs $15.00 per million). Updated October 21, 2026; 111 benchmarks have a result for both.

Replace Gemini 3.8 Flash Replace Kimi K3
Gemini 3.8 FlashGoogle
Coding
61.6
Overall
69.0
Output $/M
$3.75
61
Kimi K3Moonshot AI
Coding
62.3
Overall
66.5
Output $/M
$15.00
50
vs boards won of 111 shared

Gemini 3.8 Flash or Kimi K3: which to pick

Gemini 3.8 Flash and Kimi K3 are level for coding, 61.6 against 62.3 on the Vibe Coding Index, so price and context decide.

Pick Gemini 3.8 Flash for output tokens 4.0x cheaper ($3.75 vs $15.00 per million) and work measured by LiveBench coding or AA Coding Index, where it leads. Pick Kimi K3 for a lower cost per solved task ($0.81 against $2.74) and work measured by LMArena WebDev or SWE-bench Verified (Vals AI), where it leads.

Model AGemini 3.8 FlashGoogle61.6/1006 independent benchmarks$2.74 per solved taskModel BKimi K3Moonshot AI62.3/1006 independent benchmarks$0.81 per solved task
Shared results
111
Gemini 3.8 Flash only
17
Kimi K3 only
6
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 3.8 Flash 61.6 from 6 benchmarks, Kimi K3 62.3 from 6 benchmarks.
Widest category
Math, 11.7 points apart on average over 6 shared benchmarks.
Widest single gap
LiveBench browser: Gemini 3.8 Flash 64.0% against Kimi K3 16.0%, 48.0 points apart on a 0 to 100 scale.
Output price
Gemini 3.8 Flash $3.75 against Kimi K3 $15.00 per million output tokens, 4.0x 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 3.8 FlashBarsKimi K3Ties
Coding1212
Intelligence02
Agentic work813
Reasoning121
Math33
Knowledge64
Web design08
General207
All shared boards6150
CodingLike-for-likeSix of these boards feed the Vibe Coding Index1212
Coding benchmarks for Gemini 3.8 Flash and Kimi K3
BenchmarkGemini 3.8 FlashKimi K3Leader
LiveBench coding82.2%59.3%Gemini 3.8 Flash
LMArena WebDev15851655Kimi K3
AA Coding Index76.376.2Gemini 3.8 Flash
LiveCodeBench (Vals AI)89.5%87.2%Gemini 3.8 Flash
SWE-bench Verified (Vals AI)80.0%93.4%Kimi K3
Terminal-Bench 4 (Vals AI)19.7%17.2%Gemini 3.8 Flash
Vibe Code Bench (Vals AI)78.7%85.0%Kimi K3
Code migration (Vals AI)36.5%16.1%Gemini 3.8 Flash
ProgramBench (Vals AI)1.0%2.0%Kimi K3
IOI (Vals AI)56.9%48.9%Gemini 3.8 Flash
Terminal-Bench 2.1 (Vals AI)81.3%80.9%Gemini 3.8 Flash
SWE-bench Multimodal (Vals AI)40.8%44.2%Kimi K3
Vibe Code Bench 1-100 (Vals AI)18.8%18.2%Gemini 3.8 Flash
SkillsBench (Vals AI)58.0%Not measuredOnly Gemini 3.8 Flash measured
CyberBench (Vals AI)43.8%75.2%Kimi K3
SciCode56.6%59.5%Kimi K3
FrontierCode41.2%44.2%Kimi K3
FrontierSWE19.6%25.9%Kimi K3
WeirdML84.8%82.6%Gemini 3.8 Flash
ALE-Bench1269.71524.5Kimi K3
LMArena Text: coding15081508Gemini 3.8 Flash
LMArena Text: software and IT15071499Gemini 3.8 Flash
LMArena WebDev: React15771670Kimi K3
LMArena WebDev: HTML16161644Kimi K3
LMArena WebDev: from an image15931579Gemini 3.8 Flash
SciCode12.3%Not measuredOnly Gemini 3.8 Flash measured
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 Gemini 3.8 Flash and Kimi K3
BenchmarkGemini 3.8 FlashKimi K3Leader
AA Intelligence Index40.943.6Kimi K3
Epoch Capabilities Index157.4157.9Kimi K3
Agentic workLike-for-likePublic boards, shown as published, kept out of the index813
Agentic work benchmarks for Gemini 3.8 Flash and Kimi K3
BenchmarkGemini 3.8 FlashKimi K3Leader
AA Agentic Index40.250.0Kimi K3
Design Arena: agents building web apps12381295Kimi K3
Design Arena: agents building full-stack apps12441315Kimi K3
Design Arena: agents building mobile apps12361269Kimi K3
Design Arena: agents building Android apps10741250Kimi K3
Design Arena: agents building Godot games12731223Gemini 3.8 Flash
Design Arena: agents building gamesNot measured1252Only Kimi K3 measured
Design Arena: HTML slidesNot measured1256Only Kimi K3 measured
Design Arena: PowerPoint slides11771273Kimi K3
Design Arena: spreadsheets10561058Kimi K3
LiveBench research79.1%79.0%Gemini 3.8 Flash
LiveBench browser64.0%16.0%Gemini 3.8 Flash
LiveBench long agentic30.1%42.8%Kimi K3
SREBench (Vals AI)11.1%Not measuredOnly Gemini 3.8 Flash measured
Finance Agent (Vals AI)61.4%53.1%Gemini 3.8 Flash
Tax Agent Bench (Vals AI)32.1%33.4%Kimi K3
Legal Research Bench (Vals AI)38.9%46.1%Kimi K3
Harvey's Legal Agent Benchmark (Vals AI)12.1%12.9%Kimi K3
Excel Modeling Benchmark (Vals AI)72.2%66.7%Gemini 3.8 Flash
BioMysteryBench (Vals AI)62.2%72.6%Kimi K3
Terminal-Bench-Science (Vals AI)8.6%1.4%Gemini 3.8 Flash
Vals Science Index35.8%Not measuredOnly Gemini 3.8 Flash measured
Time Horizon Index (Vals AI)18.8%10.5%Gemini 3.8 Flash
CUA-bench (Vals AI)4.2%Not measuredOnly Gemini 3.8 Flash measured
APEX-Agents64.3%47.8%Gemini 3.8 Flash
Remote Labor Index5.8%Not measuredOnly Gemini 3.8 Flash measured
PostTrainBenchNot measured32.0%Only Kimi K3 measured
Vending-Bench 2$5,094$5,165Kimi K3
Neither model has a result on the other 13 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index121
Reasoning benchmarks for Gemini 3.8 Flash and Kimi K3
BenchmarkGemini 3.8 FlashKimi K3Leader
GPQA Diamond94.4%92.9%Gemini 3.8 Flash
LiveBench reasoning89.3%90.7%Kimi K3
MysteryMechanism (Vals AI)36.5%Not measuredOnly Gemini 3.8 Flash measured
MMMU Pro (Vals AI)89.1%88.2%Gemini 3.8 Flash
GPQA Diamond (Artificial Analysis run)95.3%93.5%Gemini 3.8 Flash
Humanity's Last Exam (Artificial Analysis run)47.8%46.9%Gemini 3.8 Flash
GPQA Diamond (Epoch AI run)95.4%93.1%Gemini 3.8 Flash
ARC-AGI98.5%94.5%Gemini 3.8 Flash
ARC-AGI-289.2%60.4%Gemini 3.8 Flash
SimpleBenchNot measured60.7%Only Kimi K3 measured
Chess puzzles58.0%37.0%Gemini 3.8 Flash
Mystery game puzzles47.0%26.0%Gemini 3.8 Flash
LMArena Text: hard prompts15091495Gemini 3.8 Flash
LMArena Text: expert15261520Gemini 3.8 Flash
LMArena Text: hard prompts in English15041496Gemini 3.8 Flash
LMArena Vision: diagrams1345Not measuredOnly Gemini 3.8 Flash measured
LMArena Vision: homework1351Not measuredOnly Gemini 3.8 Flash measured
Neither model has a result on the other 5 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index33
Math benchmarks for Gemini 3.8 Flash and Kimi K3
BenchmarkGemini 3.8 FlashKimi K3Leader
OTIS Mock AIME98.9%97.2%Gemini 3.8 Flash
FrontierMath, tiers 1 to 368.4%72.2%Kimi K3
FrontierMath, tier 422.0%39.0%Kimi K3
ProofBench48.0%87.0%Kimi K3
LMArena Text: math15321494Gemini 3.8 Flash
LMArena Text: mathematical work15191505Gemini 3.8 Flash
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index64
Knowledge benchmarks for Gemini 3.8 Flash and Kimi K3
BenchmarkGemini 3.8 FlashKimi K3Leader
MMLU Pro90.2%88.0%Gemini 3.8 Flash
LegalBench (Vals AI)87.0%86.2%Gemini 3.8 Flash
CorpFin (Vals AI)Not measured71.6%Only Kimi K3 measured
TaxEval (Vals AI)74.5%75.7%Kimi K3
MortgageTax (Vals AI)65.3%66.3%Kimi K3
MedCode (Vals AI)48.1%49.4%Kimi K3
SimpleQA Verified69.7%50.6%Gemini 3.8 Flash
LMArena Text: science15231508Gemini 3.8 Flash
LMArena Text: medicine and healthcare15081489Gemini 3.8 Flash
LMArena Text: legal and government14971503Kimi K3
LMArena Text: business and finance14791470Gemini 3.8 Flash
FACTS Benchmark Suite69.4%Not measuredOnly Gemini 3.8 Flash measured
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 Gemini 3.8 Flash and Kimi K3
BenchmarkGemini 3.8 FlashKimi K3Leader
Design Arena: websites13001335Kimi K3
Design Arena: UI components13191344Kimi K3
Design Arena: game dev13051359Kimi K3
Design Arena: data viz12691351Kimi K3
Design Arena: 3D12951388Kimi K3
Design Arena: all code categories13041360Kimi K3
Design Arena: SVG12951302Kimi K3
Design Arena: ASCII art1265Not measuredOnly Gemini 3.8 Flash measured
LiveBench design72.7%77.2%Kimi K3
GeneralLike-for-likePublic boards, shown as published, kept out of the index207
General benchmarks for Gemini 3.8 Flash and Kimi K3
BenchmarkGemini 3.8 FlashKimi K3Leader
Vals Index56.0%53.0%Gemini 3.8 Flash
LiveBench data analysis54.0%78.7%Kimi K3
Public Benefits Bench (Vals AI)65.3%68.2%Kimi K3
SAGE (Vals AI)35.1%52.8%Kimi K3
MedScribe (Vals AI)84.5%88.0%Kimi K3
Vals Multimodal IndexNot measured73.4%Only Kimi K3 measured
Long context reasoning (AA-LCR)84.0%88.7%Kimi K3
DTBench95.7%91.2%Gemini 3.8 Flash
LMCA52.9%52.7%Gemini 3.8 Flash
Surface Evolver Bench76.9%95.0%Kimi K3
Furniture assembly31.7%34.2%Kimi K3
LMArena Text14991476Gemini 3.8 Flash
LMArena Text: instruction following14921485Gemini 3.8 Flash
LMArena Text: creative writing14881457Gemini 3.8 Flash
LMArena Vision1315Not measuredOnly Gemini 3.8 Flash measured
LMArena Text: English14981481Gemini 3.8 Flash
LMArena Text: non-English14911466Gemini 3.8 Flash
LMArena Text: Chinese15531529Gemini 3.8 Flash
LMArena Text: German15011490Gemini 3.8 Flash
LMArena Text: Japanese15071494Gemini 3.8 Flash
LMArena Text: Korean14641460Gemini 3.8 Flash
LMArena Text: French14961492Gemini 3.8 Flash
LMArena Text: Spanish14871472Gemini 3.8 Flash
LMArena Text: Russian15141485Gemini 3.8 Flash
LMArena Text: Polish14971458Gemini 3.8 Flash
LMArena Text: multi-turn15031489Gemini 3.8 Flash
LMArena Text: longer queries15071494Gemini 3.8 Flash
LMArena Text: writing and language14941471Gemini 3.8 Flash
LMArena Text: entertainment and media14801453Gemini 3.8 Flash
LMArena Vision: English1318Not measuredOnly Gemini 3.8 Flash measured
LMArena Vision: Chinese1416Not measuredOnly Gemini 3.8 Flash measured
LMArena Vision: reading text1333Not measuredOnly Gemini 3.8 Flash measured
LMArena Vision: humor1321Not measuredOnly Gemini 3.8 Flash measured
LMArena Vision: creative writing1299Not measuredOnly Gemini 3.8 Flash measured
Neither model has a result on the other 12 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Gemini 3.8 Flash and Kimi K3
BenchmarkGemini 3.8 FlashKimi 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 3.8 Flash$0.0026Fits in one request
Kimi K3$0.011Fits in one request

Gemini 3.8 Flash costs less

Repository review

50K input + 3K output tokens

Gemini 3.8 Flash$0.049Fits in one request
Kimi K3$0.195Fits in one request

Gemini 3.8 Flash costs less

Agent loop

200K input + 10K output tokens

Gemini 3.8 Flash$0.188Fits in one request
Kimi K3$0.750Fits in one request

Gemini 3.8 Flash 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 3.8 Flash and Kimi K3
SpecificationGoogle Gemini 3.8 FlashMoonshotAI Kimi K3
MakerGoogleMoonshot AI
WeightsNo public weights listedOpen weights
ReleasedSeptember 2, 2026July 16, 2026
Context window1.05M1.05M
Input $/Mtok$0.75$3.00
Output $/Mtok$3.75$15.00
Coding index61.662.3
Independent measurements behind it66
Cost per solved task$2.74$0.81
SourceOpenRouter listingOpenRouter listing

Questions

Is Gemini 3.8 Flash better than Kimi K3 for coding?

Gemini 3.8 Flash and Kimi K3 are level for coding, 61.6 against 62.3 on the Vibe Coding Index, so price and context decide. Pick Gemini 3.8 Flash for output tokens 4.0x cheaper ($3.75 vs $15.00 per million) and work measured by LiveBench coding or AA Coding Index, where it leads. Pick Kimi K3 for a lower cost per solved task ($0.81 against $2.74) and work measured by LMArena WebDev or SWE-bench Verified (Vals AI), where it leads.

Which is cheaper, Gemini 3.8 Flash or Kimi K3?

Gemini 3.8 Flash costs $0.75 per million input tokens and $3.75 per million output tokens; Kimi K3 costs $3.00 and $15.00. Per solved SWE-bench Verified task, Gemini 3.8 Flash costs $2.74 and Kimi K3 $0.81.

Which has the bigger context window, Gemini 3.8 Flash or Kimi K3?

Neither: both take 1.05M tokens.

How many benchmarks is the Gemini 3.8 Flash vs Kimi K3 comparison based on?

111 benchmarks have a result for both models as of October 21, 2026. Gemini 3.8 Flash's coding index rests on 6 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-3-8-flash-vs-kimi-k3" title="Gemini 3.8 Flash 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-3-8-flash-vs-kimi-k3">LLMBenchmark.ai, Gemini 3.8 Flash vs Kimi K3</a></p>