Skip to main content
BenchLM
Advertise

Model comparison

Gemini 3 Flash vs Kimi K2.5

Data verified

Head-to-head evidence from 18 shared benchmark results across 7 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

60.49/100
Margin
0.8pts
← winning
Kimi K2.5
Moonshot AI
59.66/100
0 category wins1 category wins

Public leaderboard positions: Gemini 3 Flash #49 (Supported); Kimi K2.5 #54 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. Gemini 3 Flash and Kimi K2.5 share 18 comparable benchmark results. 1 of 8 categories are comparable. 1 result is unique to Gemini 3 Flash; 45 to Kimi K2.5.

Updated July 22, 2026
Shared results
18
Gemini 3 Flash only
1
Kimi K2.5 only
45
Comparable categories
1 / 8

Pick Gemini 3 Flash if you want the stronger benchmark profile. Kimi K2.5 only becomes the better choice if mathematics is the priority.

Confidence note. This is a partial-evidence comparison with 18 shared benchmark results across 7 evidence categories; 1 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

Gemini 3 Flash has the cleaner BenchAlign overall profile here, landing at 60.49 versus 59.66. It is a real lead, but still close enough that category-level strengths matter more than the headline number.

Gemini 3 Flash gives you the larger context window at 1M, compared with 256K for Kimi K2.5.

Category breakdown

Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.

Category scores and score margins for Gemini 3 Flash and Kimi K2.5
CategoryGemini 3 FlashΔKimi K2.5
MathGemini 3 Flash27.8Margin→ 32.8Kimi K2.560.6
AgenticGemini 3 FlashNot measuredMarginNo overlapKimi K2.555.0
CodingGemini 3 FlashNot measuredMarginNo overlapKimi K2.559.4
ReasoningGemini 3 FlashNot measuredMarginNo overlapKimi K2.561.0
KnowledgeGemini 3 FlashNot measuredMarginNo overlapKimi K2.556.9
MultilingualGemini 3 FlashNot measuredMarginNo overlapKimi K2.582.3
MultimodalGemini 3 FlashNot measuredMarginNo overlapKimi K2.578.5
Inst. FollowingGemini 3 FlashNot measuredMarginNo overlapKimi K2.593.9

Decisive benchmark drivers

The largest measured benchmark gaps in this matchup, with exact reported values.

More
A · Gemini 3 FlashB · Kimi K2.5

  1. FrontierMath v2 (Tiers 1-3)

    Math
    Source ↗
    A 35.640%B 27.900%
    Winner: Gemini 3 FlashΔ 7.7
    FrontierMath v2 (Tiers 1-3): Gemini 3 Flash scored 35.640%; Kimi K2.5 scored 27.900%. Gemini 3 Flash wins this benchmark.

  2. FrontierMath v2 (Tier 4)

    Math
    Source ↗
    A 4.167%B 4.200%
    Winner: Kimi K2.5Δ 0
    FrontierMath v2 (Tier 4): Gemini 3 Flash scored 4.167%; Kimi K2.5 scored 4.200%. Kimi K2.5 wins this benchmark.

Operational comparison

Runtime and commercial metrics are compared only when both models have a complete sourced value.

MetricGemini 3 FlashKimi K2.5Comparison
Input / output priceUSD per 1M tokensGemini 3 Flash$0.5 input / $3 outputKimi K2.5$0.6 input / $3 outputGemini 3 Flash has the lower combined listed price.
Generation speedtokens per secondGemini 3 Flash159 tok/sKimi K2.545 tok/sGemini 3 Flash has the higher measured throughput.
First-answer latencyseconds to first tokenGemini 3 Flash1.19 sKimi K2.52.38 sGemini 3 Flash reaches the first token sooner.
Context windowmaximum listed tokensGemini 3 Flash1MKimi K2.5256KGemini 3 Flash lists the larger context window.

Benchmark Deep Dive

Agentic
BenchmarkGemini 3 FlashKimi K2.5Result
Claw-EvalSource 49.2%52.3%Kimi K2.5 leads
τ²-bench resultsSource 43.3%95.9%Kimi K2.5 leads
Gert LabsSource 56.63%45.88%Gemini 3 Flash leads
JobBenchSource 11.4%8.7%Gemini 3 Flash leads
Terminal-Bench 2.0Source —50.8%Not comparable
BrowseCompSource —60.6%Not comparable
QwenClawBenchSource —54.3%Not comparable
τ³-bench resultsSource —65.7%Not comparable
DeepSearchQASource —77.1%Not comparable
DeepPlanningSource —14.4%Not comparable
ToolathlonSource —27.8%Not comparable
MCP AtlasSource —29.5%Not comparable
MCP-TasksSource —59.1%Not comparable
WideResearchSource —72.7%Not comparable
APEX-Agents-AASource —11.5%Not comparable
ResearchClawBenchSource —14.0%Not comparable
AA Agentic IndexSource —21.7%Not comparable
GDPval-AASource —25.4%Not comparable
GDPval-AASource —1009Not comparable
Coding
BenchmarkGemini 3 FlashKimi K2.5Result
Vibe Code BenchSource 20.20%—Not comparable
AA-SciCodeSource 49.9%49.0%Gemini 3 Flash leads
SWE-bench VerifiedSource —76.8%Not comparable
SWE-bench Verified*Source —70.8%Not comparable
LiveCodeBench v6Source —85.0%Not comparable
SWE-bench ProSource —50.7%Not comparable
SWE MultilingualSource —73%Not comparable
SWE-RebenchSource —58.5%Not comparable
React Native EvalsSource —77.2%Not comparable
SciCodeSource —48.7%Not comparable
AA Coding IndexSource —46.8%Not comparable
Reasoning
BenchmarkGemini 3 FlashKimi K2.5Result
AA-LCRSource 48.0%65.3%Kimi K2.5 leads
CritPtSource 1.4%3.1%Kimi K2.5 leads
LongBench v2Source —61%Not comparable
Knowledge
BenchmarkGemini 3 FlashKimi K2.5Result
Artificial Analysis Intelligence IndexSource 27.4%35.4%Kimi K2.5 leads
AA-GPQA DiamondSource 81.2%87.9%Kimi K2.5 leads
AA-HLESource 14.1%29.4%Kimi K2.5 leads
AA-Omniscience IndexSource -3.6%-8.1%Gemini 3 Flash leads
AA-Omniscience AccuracySource 45.5%34.3%Gemini 3 Flash leads
AA-Omniscience Hallucination RateSource 90.2%64.6%Kimi K2.5 leads
GPQASource —87.6%Not comparable
GPQA-DSource —87.6%Not comparable
SuperGPQASource —69.2%Not comparable
MMLU-ProSource —87.1%Not comparable
MMLU-Pro (Arcee)Source —87.1%Not comparable
HLESource —30.1%Not comparable
MathKimi K2.5 wins
BenchmarkGemini 3 FlashKimi K2.5Result
FrontierMath v2 (Tiers 1-3)Source 35.640%27.900%Gemini 3 Flash leads
FrontierMath v2 (Tier 4)Source 4.167%4.200%Kimi K2.5 leads
AIME 2025Source —96.1%Not comparable
AIME26Source —95.8%Not comparable
AIME25 (Arcee)Source —96.3%Not comparable
HMMT Feb 2025Source —95.4%Not comparable
HMMT Nov 2025Source —91.1%Not comparable
HMMT Feb 2026Source —87.1%Not comparable
MMAnswerBenchSource —81.8%Not comparable
Multilingual
BenchmarkGemini 3 FlashKimi K2.5Result
MMLU-ProXSource —82.3%Not comparable
NOVA-63Source —56.0%Not comparable
Multimodal
BenchmarkGemini 3 FlashKimi K2.5Result
AA-MMMU-ProSource 78.6%75.4%Gemini 3 Flash leads
Design Arena WebsiteSource 12261279Kimi K2.5 leads
MMMU-ProSource —78.5%Not comparable
Video-MMESource —87.4%Not comparable
MMVUSource —80.4%Not comparable
VideoMMMUSource —86.6%Not comparable
Inst. Following
BenchmarkGemini 3 FlashKimi K2.5Result
AA-IFBenchSource 55.1%70.2%Kimi K2.5 leads
IFEvalSource —93.9%Not comparable
Frequently Asked Questions (2)

Which is better, Gemini 3 Flash or Kimi K2.5?

Gemini 3 Flash is ahead on BenchLM's BenchAlign leaderboard, 60.49 to 59.66. The biggest single separator in this matchup is FrontierMath v2 (Tiers 1-3), where the scores are 35.640% and 27.900%.

Which is better for math, Gemini 3 Flash or Kimi K2.5?

Kimi K2.5 has the edge for math in this comparison, averaging 60.6 versus 27.8. Inside this category, FrontierMath v2 (Tiers 1-3) is the benchmark that creates the most daylight between them.

Self-host vs API cost

Estimates at 50,000 req/day · 1000 tokens/req average.

Gemini 3 Flash
API / mo$2,625
Self-host / moNot listed
Break-even—
Proprietary model — self-hosting not applicable.
Kimi K2.5
API / mo$2,700
Self-host / mo$5,221
Break-even132M/day
Model the full break-even

Related Comparisons

Explore More

Google Compare Pricing Methodology Find Your Best LLM Overall Rankings
Last updated: July 22, 2026

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.