BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases4

Gemini 3.6 Flash vs Kimi K3.1 Pro

Full profiles: Gemini 3.6 Flash (Google) and Kimi K3.1 Pro (Moonshot AI): every benchmark, price and rank.

Updated · Data from 10 sources

As of 2 Nov 2026, Gemini 3.6 Flash and Kimi K3.1 Pro are neck and neck on quality (61.3 vs 61.5 on the BenchLeader Index, inside the ±4.7 margin), and Gemini 3.6 Flash is 2.7× cheaper at $1.50 vs $4.00 per 1M tokens blended. Choose Gemini 3.6 Flash for reasoning (64 vs 60) and fast first answers (16.4 s vs 68.3 s), where it also streams 3×+ faster; choose Kimi K3.1 Pro for maths (74 vs 59), agents & tools (68 vs 59), and long context (68 vs 65). Gemini 3.6 Flash's current price runs until 31 Dec 2026, then $3.00.

Key facts (as of 2 Nov 2026)

  • Gemini 3.6 Flash scores 61.3 on the BenchLeader Index and Kimi K3.1 Pro scores 61.5 (BenchLeader, 2 Nov 2026).
  • Gemini 3.6 Flash costs $1.50 per 1M tokens blended until 31 Dec 2026, then $3.00 (https://ai.google.dev/gemini-api/docs/pricing, checked 20 Oct 2026).
  • Kimi K3.1 Pro costs $4.00 per 1M tokens blended (list price, 2 Nov 2026).
  • Gemini 3.6 Flash streams 3×+ faster than Kimi K3.1 Pro (Artificial Analysis, 7-day median, 2 Nov 2026).
  • Of the 10 benchmarks both models have, Kimi K3.1 Pro leads 10 (BenchLeader, 2 Nov 2026).
Gemini 3.6 Flash benchmarks high Google Kimi K3.1 Pro benchmarks Moonshot AIAdd a third model

Which to pick for…

The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.

Use casePickWhyConfidence
CodingKimi K3.1 Pro60 vs 57 category score; SciCode 60.8% vs 53.4%Medium
Agents & tool useKimi K3.1 Pro68 vs 59 category score; AutomationBench 73.3% vs 53.0%High
MathsKimi K3.1 Pro74 vs 59 category scoreHigh
ReasoningGemini 3.6 Flash64 vs 60 category score; Humanity's Last Exam (AA) 40.8% vs 52.5%Medium
KnowledgeEither64 vs 63 category scoreEither
Long documentsKimi K3.1 Pro68 vs 65 category score; AA-LCR v1.1 86.3% vs 80.0%Medium
Images and other mediaKimi K3.1 Pro67 vs 65 category score; MMMU-Pro 83.9% vs 83.2%Medium
Chat quality (human preference)–Only Gemini 3.6 Flash has an Arena rating (67 category score)Not comparable
Lowest costGemini 3.6 Flash$1.50 vs $4.00 per 1M blended; agentic coding session $0.060 vs $0.160; from 1 Jan 2027: $3.00 vs $4.00–
Fastest first answerGemini 3.6 Flash16.4 s (Gemini 3.6 Flash (high)) vs 68.3 s (Kimi K3.1 Pro) to the first answer token–
High volume on a budgetNeither flagshipA smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs–

Key differences

  • Gemini 3.6 Flash (high) and Kimi K3.1 Pro are neck and neck on quality (61.3 vs 61.5, within the ±4.7 margin).
  • Gemini 3.6 Flash (high) is stronger in human preference, knowledge, reasoning.
  • Kimi K3.1 Pro is stronger in agents & tools, coding, long context, maths, multimodal, composite.
  • Gemini 3.6 Flash (high) is 2.7× cheaper ($1.50 vs $4.00 per 1M blended) until 31 Dec 2026, then $3.00 vs $4.00 from 1 Jan 2027.
  • Gemini 3.6 Flash (high) streams 6.4× faster (197 vs 31 tokens per second).

At a glance

SpecGemini 3.6 FlashKimi K3.1 Pro
DeveloperGoogleMoonshot AI
Released21 Jul 202623 Oct 2026
Context window (tokens)1,048,5761,048,576
Compared atGemini 3.6 Flash (high)Kimi K3.1 Pro
Open weightsNoNo
Price in / out / cached input per 1M$0.750 / $3.75 / $0.075$2.00 / $10.00 / $0.200
Announced price change$1.50 / $7.50 / $0.150 from 1 Jan 2027None announced

Limits, availability and announced prices from the vendors' own pages: ai.google.dev pricing (checked 20 Oct 2026). Everything else is from the day's data.

Category by category

Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.

CategoryGemini 3.6 Flash (high)Kimi K3.1 ProLeader
Reasoning64.059.7Gemini 3.6 Flash
Coding56.960.0Kimi K3.1 Pro
Agents & tools59.367.9Kimi K3.1 Pro
Maths59.473.7Kimi K3.1 Pro
Knowledge63.862.9level
Human preference67.4–not comparable
Multimodal65.266.6Kimi K3.1 Pro
Long context64.868.1Kimi K3.1 Pro
Composite–56.5not comparable

On one chart

Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.

2550 avg75ReasoningCodingAgents & toolsMathsKnowledgeLong contextMultimodalGemini 3.6 Flash, Reasoning: 64.0Gemini 3.6 Flash, Coding: 56.9Gemini 3.6 Flash, Agents & tools: 59.3Gemini 3.6 Flash, Maths: 59.4Gemini 3.6 Flash, Knowledge: 63.8Gemini 3.6 Flash, Long context: 64.8Gemini 3.6 Flash, Multimodal: 65.2Kimi K3.1 Pro, Reasoning: 59.7Kimi K3.1 Pro, Coding: 60.0Kimi K3.1 Pro, Agents & tools: 67.9Kimi K3.1 Pro, Maths: 73.7Kimi K3.1 Pro, Knowledge: 62.9Kimi K3.1 Pro, Long context: 68.1Kimi K3.1 Pro, Multimodal: 66.6

Not drawn, measured for fewer than two of these models: Instruction following, Human preference.

ModelIndex

Where sources disagree

Across the 10 benchmarks both have, Gemini 3.6 Flash leads 0, Kimi K3.1 Pro leads 10.

Overall indexGemini 3.6 Flash (high)Kimi K3.1 ProLeader
BenchLeader Index61.361.5neck and neck

Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).

Price, speed and latency

List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.

MeasureGemini 3.6 Flash (high)Kimi K3.1 Pro
Blended price per 1M tokens, until 31 Dec 2026$1.50$4.00
Input / output per 1M$0.750 / $3.75$2.00 / $10.00
Blended price per 1M from 1 Jan 2027$3.00$4.00
Input / output per 1M from 1 Jan 2027$1.50 / $7.50$2.00 / $10.00
Cached input per 1M$0.075$0.200
Price per task: chat reply$0.0014$0.0038
Price per task: summarize a 30-page report$0.011$0.030
Price per task: code edit$0.010$0.027
Price per task: agentic coding session$0.060$0.160
Price per task: structured extraction$0.0023$0.0060
Agentic coding session, input mostly cached$0.030$0.079
Cost per benchmark task (AA Index)–$3.68
Output tokens per task (AA Index)–72k
Of which reasoning–44k
Time per task (AA Index)–34 min
Output speed197 tok/s31 tok/s
Time to first answer16.4 s68.3 s
Time to a complete answer18.9 s84.5 s

Each model is compared at its best-ranked reasoning-effort setting. Gemini 3.6 Flash (high) scores 61.3 and takes 16 s to its first answer token; Gemini 3.6 Flash scores 49.5 in 1 s. Kimi K3.1 Pro scores 61.5 and takes 68 s to its first answer token. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.

Every benchmark

122 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.

MetricPublisherGemini 3.6 Flash (high)Kimi K3.1 Pro
BenchLeader IndexBenchLeader61.361.5
Agents & tools scoreBenchLeader59.367.9
Coding scoreBenchLeader56.960.0
Human preference scoreBenchLeader67.4–
Knowledge scoreBenchLeader63.862.9
Long context scoreBenchLeader64.868.1
Maths scoreBenchLeader59.473.7
Multimodal scoreBenchLeader65.266.6
Reasoning scoreBenchLeader64.059.7
Composite scoreBenchLeader–56.5
Blended price $/MBenchLeader$1.50$4.00
Output speedBenchLeader197 tok/s31 tok/s
Time to first answerBenchLeader16.4 s68.3 s
Context windowBenchLeader1,048,576 tokens1,048,576 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub94.1%–
Arena Hard PromptsArena (LMArena)1501–
LiveBench 2.0 ReasoningLiveBench–91.1%
LiveBench 1.0 ReasoningLiveBench85.2%–
GPQA Diamond (AA)Artificial Analysis92.8%–
Humanity's Last Exam (AA)Artificial Analysis40.8%52.5%
HLE-Diamond (AA)Artificial Analysis–37.3%
GPQA Diamond (Vals)Vals AI93.4%–
ARC-AGI-1ARC Prize91.2%–
ARC-AGI-2ARC Prize60.4%–
CritPtArtificial Analysis10.6%24.6%
CritPt v2 hard subset (AA)Artificial Analysis–45.0%
MysteryMechanismVals AI–26.1%†
Chess PuzzlesEpoch AI Benchmarking Hub40.0%–
Mystery Game PuzzlesEpoch AI Benchmarking Hub30.0%–
LMCAEpoch AI Benchmarking Hub44.9%–
DTBenchEpoch AI Benchmarking Hub95.5%–
ForecastBenchForecastBench65.9%†–
Coding
SciCodeSciCode53.4%60.8%
WeirdMLWeirdML56.1%–
FrontierCodeCognition34.4%†–
Arena CodingArena (LMArena)1518–
Arena WebDevArena (LMArena)15381695†
LiveBench 2.0 CodingLiveBench–81.8%
LiveBench 1.0 CodingLiveBench77.9%–
Show all 122 benchmarks
MetricPublisherGemini 3.6 Flash (high)Kimi K3.1 Pro
Coding
LiveCodeBenchVals AI88.1%–
IOIVals AI35.1%81.9%†
SWE-bench (Vals)Vals AI79.6%–
Code MigrationVals AI30.9%–
ProgramBenchVals AI0.0%–
Vibe Code Bench v1.1Vals AI64.0%84.9%†
DeepSWE v1.1Epoch AI Benchmarking Hub46.7%–
ALE-BenchEpoch AI Benchmarking Hub715.51932.8†
Terminal-Bench 4.0 (AA)Artificial Analysis7.1%67.2%
Terminal-Bench 2.1 (AA)Artificial Analysis77.5%–
Startrise FrontendStartrise–82.5%
GameBench 2 OverallDuelLab GameBench 252.5†–
GameBench 2 by reasoning settingDuelLab GameBench 264.680.8†
FrontierCode 1.1 (AA)Artificial Analysis–47.1%
FrontierCode 2.0 (AA)Artificial Analysis–25.4%
Agents & tools
APEX-AgentsMercor46.9%†–
Arena AgentArena (LMArena)–6†
LiveBench Long AgenticLiveBench–39.1%
LiveBench BrowserLiveBench–75.9%
LiveBench Agentic CodingLiveBench43.4%–
Terminal-Bench 2.1 (Vals)Vals AI73.8%–
GDPval-AA v2.1, normalised scoreArtificial Analysis–67.3%
GDPval-AA v2.1Artificial Analysis–1858
τ³-Banking (AA)Artificial Analysis29.9%–
Analyst Agent (AA)Artificial Analysis–38.8%
CyberBenchVals AI45.4%–
Finance Agent v2Vals AI56.3%–
Finance Agent v2.1Vals AI–61.5%†
Harvey's Legal Agent Benchmark v1.1Vals AI–11.7%†
Harvey's Legal Agent Benchmark v1Vals AI3.3%–
Legal Research BenchVals AI25.0%37.5%†
Tax Agent BenchVals AI18.0%24.9%†
Terminal-Bench 4.0 (Vals)Vals AI–63.6%†
Terminal-Bench ScienceVals AI4.3%31.4%†
AutomationBenchArtificial Analysis53.0%73.3%
GDP.pdf (AA)Artificial Analysis17.4%31.4%
AA-Briefcase v1.1Artificial Analysis–1676
Terminal-Bench-Science 0.1 (AA)Artificial Analysis–34.3%
CWE-Bench (AA)Artificial Analysis–60.8%
DeepSec-Bench (AA)Artificial Analysis–36.7%
CyberGym end-to-end (AA)Artificial Analysis–94.7%
FACTS SearchGoogle68.2%†–
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub59.0%–
FrontierMath Tier 4Epoch AI Benchmarking Hub21.9%–
OTIS Mock AIMEEpoch AI Benchmarking Hub94.2%–
ProofBenchVals AI36.0%†91.0%
LiveBench MathematicsLiveBench86.4%–
AIME 2026MathArena96.7%†–
HMMT February 2026MathArena89.4%†–
MathArena ApexMathArena26.0%†–
Arena MathsArena (LMArena)1504–
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub66.2%–
LiveBench 2.0 Data AnalysisLiveBench–79.1%
LiveBench ResearchLiveBench–77.9%
LiveBench 1.0 Data AnalysisLiveBench63.0%–
AA-Omniscience v1.1Artificial Analysis–29.9
MMLU-ProVals AI89.3%–
LegalBenchVals AI86.7%–
CorpFinVals AI63.3%–
TaxEvalVals AI74.9%–
BioMysteryBenchVals AI58.5%75.6%†
Excel Modeling BenchmarkVals AI65.4%–
Excel Modeling Benchmark v1.1Vals AI–70.0%†
MedCodeVals AI53.1%–
MedScribeVals AI79.7%–
Public Benefits Bench v1.1Vals AI56.6%–
GDP.pdfEpoch AI Benchmarking Hub14.0%–
AA-Omniscience v1.1: accuracyArtificial Analysis–48.9%
AA-Omniscience v1.1: non-hallucinationArtificial Analysis–62.7%
Instruction following
LiveBench LanguageLiveBench83.9%–
LiveBench Instruction FollowingLiveBench75.4%–
Arena Instruction FollowingArena (LMArena)1473–
Human preference
Arena TextArena (LMArena)1483–
Arena Creative WritingArena (LMArena)1471–
Arena Multi-turnArena (LMArena)1488–
Arena Longer QueriesArena (LMArena)1485–
Multimodal
Arena VisionArena (LMArena)1299–
LiveBench DesignLiveBench–77.8%
MMMU-ProArtificial Analysis83.2%83.9%
MMMU-Pro (Vals)Vals AI88.4%–
MortgageTaxVals AI67.6%–
Blueprint-Bench 2Epoch AI Benchmarking Hub31.2%†–
AA-3DArtificial Analysis–1110
Long context
AA-LCR v1.1Artificial Analysis80.0%86.3%
Arena DocumentArena (LMArena)1453–
MLCRArtificial Analysis10.0%38.3%
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub154.7†–
LiveBench 2.0LiveBench–71.3%
LiveBench 1.0LiveBench73.6%–
AA Intelligence Index v5.0Artificial Analysis–46
AA Intelligence Index v4.3.2Artificial Analysis34.055.1
AA Progress Index v1Artificial Analysis305.7340.7
Vals IndexVals AI–62.5†
Vals Finance IndexVals AI–56.0%†
Vals Science IndexVals AI–44.4%†
Vals Multimodal IndexVals AI65.1%–
AA Cyber IndexArtificial Analysis–64.1

Data as of 2026-11-02. Best configuration of each model; every score links to its source on the model pages.

DuelLab GameBench 2, release gb2-20261026-cb69502a12d9 (https://benchmarks.duellab.org/releases/gb2-20261026-cb69502a12d9/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

What changed

  • Gemini 3.6 Flash: list price moves to $1.50 / $7.50 per 1M input / output tokens (announced).
  • Kimi K3.1 Pro: 1 score revision.
  • Kimi K3.1 Pro: BenchLeader Index 59.3 → 61.3.
  • Kimi K3.1 Pro: 3 new results (SciCode, ProofBench, Startrise Frontend).
  • Gemini 3.6 Flash (high): BenchLeader Index 61.8 → 61.2.
  • Kimi K3.1 Pro: BenchLeader Index 61.8 → 59.3.
  • Gemini 3.6 Flash (high): 1 new result (AA Progress Index v1).
  • Kimi K3.1 Pro: 5 new results (AA Intelligence Index v5.0, AA Progress Index v1, HLE-Diamond (AA)…); 2 score revisions; Kimi K3.1 Pro cost per task moved from $2.11/M to $3.68/M (74% up).
  • Kimi K3.1 Pro: BenchLeader Index 62.3 → 61.8.
  • Kimi K3.1 Pro: 1 new result (MLCR); 3 score revisions.

Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs

Gemini 3.6 Flash vs Kimi K3.1 Pro: questions

Is Gemini 3.6 Flash better than Kimi K3.1 Pro?
Gemini 3.6 Flash (high) and Kimi K3.1 Pro are neck and neck on quality (61.3 vs 61.5, within the ±4.7 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 2 Nov 2026, so check the category scores for your use.
Gemini 3.6 Flash vs Kimi K3.1 Pro: which is cheaper, and will that change?
Gemini 3.6 Flash is cheaper today: $1.50 against $4.00 per million tokens, blended at three input tokens per output token. Gemini 3.6 Flash's current price runs until 31 Dec 2026, then $3.00.
Is Gemini 3.6 Flash or Kimi K3.1 Pro better for coding?
Kimi K3.1 Pro scores higher in coding (60 vs 57 on the category index, where 50 is average). On SciCode, the top shared coding benchmark, it is 53.4% for Gemini 3.6 Flash against 60.8% for Kimi K3.1 Pro.
Which is faster, Gemini 3.6 Flash or Kimi K3.1 Pro?
Gemini 3.6 Flash streams 3×+ faster (197 vs 31 output tokens per second, Artificial Analysis 7-day median).
Is Gemini 3.6 Flash or Kimi K3.1 Pro better for agents and tool use?
Kimi K3.1 Pro scores higher in agentic tasks (68 vs 59 on the category index, where 50 is average).
More questions (4)
Which is better for long documents?
Kimi K3.1 Pro scores higher on long-context benchmarks (68 vs 65). Both accept 1,048,576 tokens of context.
Which has lower latency?
Gemini 3.6 Flash answers first: 16.4 s against 68.3 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
Do independent benchmarks agree on Gemini 3.6 Flash vs Kimi K3.1 Pro?
Across the 10 benchmarks both have, Gemini 3.6 Flash leads 0, Kimi K3.1 Pro leads 10.
When was Kimi K3.1 Pro released?
Kimi K3.1 Pro was released on 23 Oct 2026.