Head-to-head comparison

Gemini 3.1 Pro Preview vs. Kimi K3

Compare published benchmark results from matching versions and cohorts, API pricing, and technical specifications. Any documented test setup differences remain visible.

shared result series: 152
distinct benchmarks: 30
latest retrieval: 2026-09-29

First modelSecond model

Specifications

General specifications

Release date, availability, context window, architecture, and published parameter counts at a glance. Undisclosed details remain marked as unavailable.

Specifications for Gemini 3.1 Pro Preview and Kimi K3
Attribute	Gemini 3.1 Pro Preview	Kimi K3
Provider	Google	Moonshot AI
Released	Feb 19, 2026	Jul 16, 2026
Availability	Preview	Active
Model type	Proprietary	Open Weights
Parameters	Not published	2.8T, 104B active
Architecture	Not published	Mixture of Experts
Context window	1,048,576 tokens	1,048,576 tokens
Knowledge cutoff	Not published	Not published
Technical sources	Model	Model

Pricing

API pricing compared

The table compares official standard API rates per 1M tokens, including published cache pricing. Batch, fast, priority, regional surcharges, tool calls, and cloud platform pricing are excluded.

API pricing for Gemini 3.1 Pro Preview and Kimi K3
Price type	Gemini 3.1 Pro Preview	Kimi K3
API input	$2 ≤200K / $4 >200K	$3
API output	$12 ≤200K / $18 >200K	$15
Cache read	$0.2 ≤200K / $0.4 >200K	$0.3
Cache write (5 min.)	Not listed	Not listed
Cache write (1 hr.)	Not listed	Not listed
Price verified	09/22/2026 Source	09/22/2026 Source

Gemini 3.1 Pro PreviewKimi K3

API inputUSD per 1M tokens, base tier

Gemini 3.1 Pro Preview$2

Kimi K3$3

API outputUSD per 1M tokens, base tier

Gemini 3.1 Pro Preview$12

Kimi K3$15

Cache readUSD per 1M tokens, base tier

Gemini 3.1 Pro Preview$0.2

Kimi K3$0.3

Key to the marks:better valuebetter, but not conclusivetie

Pricing by provider

Direct developer pricing and provider endpoint tariffs routed through OpenRouter are listed separately. Prices are in USD per 1M tokens. Context-dependent direct tiers plus regional, Flex, Priority, and other OpenRouter tariffs remain individually identifiable.

Provider pricing for Gemini 3.1 Pro Preview and Kimi K3
Model	Provider	Purchase route and tariff	Input per 1M tokens	Output per 1M tokens	Source
Gemini 3.1 Pro Preview	Google	Direct from developerUp to 200,000 context tokens	$2	$12	Source09/22/2026
	Google	Direct from developerAbove 200,000 context tokens	$4	$18	Source09/22/2026
	Google	Via OpenRoutergoogle-vertex/global	$2	$12	Source10/02/2026
	Google	Via OpenRoutergoogle-vertex/global/flex	Lowest listed price: $1	Lowest listed price: $6	Source10/02/2026
	Google	Via OpenRoutergoogle-vertex/global/priority	$3.6	$21.6	Source10/02/2026
	Google AI Studio	Via OpenRoutergoogle-ai-studio	$2	$12	Source10/02/2026
	Google AI Studio	Via OpenRoutergoogle-ai-studio/flex	Lowest listed price: $1	Lowest listed price: $6	Source10/02/2026
	Google AI Studio	Via OpenRoutergoogle-ai-studio/priority	$3.6	$21.6	Source10/02/2026
Kimi K3	Moonshot AI	Direct from developerStandard API	$3	$15	Source09/22/2026
	AkashML	Via OpenRouterakashml/fp4	$3	$15	Source10/02/2026
	Alibaba	Via OpenRouteralibaba	$3.45	$17.25	Source10/02/2026
	BaseTen	Via OpenRouterbaseten/fp8	$3	$15	Source10/02/2026
	Chutes	Via OpenRouterchutes/mxfp4	$3	$15	Source10/02/2026
	Decart	Via OpenRouterdecart/mxfp4	$3	$15	Source10/02/2026
	DeepInfra	Via OpenRouterdeepinfra/bf16	$2.85	$14.25	Source10/02/2026
	DigitalOcean	Via OpenRouterdigitalocean	$2.55	$12.95	Source10/02/2026
	Fireworks	Via OpenRouterfireworks	$3	$15	Source10/02/2026
	Fireworks	Via OpenRouterfireworks/fast	$4.5	$22.5	Source10/02/2026
	Fireworks	Via OpenRouterfireworks/us	$4.5	$22.5	Source10/02/2026
	InferenceNet	Via OpenRouterinference-net/fp4	Lowest listed price: $1.34	$13	Source10/02/2026
	Makora	Via OpenRoutermakora	$2.04	$12.75	Source10/02/2026
	Modal	Via OpenRoutermodal/mxfp4	$3	$15	Source10/02/2026
	Moonshot AI	Via OpenRoutermoonshotai/mxfp4	$3	$15	Source10/02/2026
	Morph	Via OpenRoutermorph/fp8	$1.35	$11.357	Source10/02/2026
	Parasail	Via OpenRouterparasail/fp4	$3	$15	Source10/02/2026
	Phala	Via OpenRouterphala	$2.25	$11.25	Source10/02/2026
	Relace	Via OpenRouterrelace/fp4	$1.8	Lowest listed price: $11.2	Source10/02/2026
	Sail Research	Via OpenRoutersail-research/fp4	$2.8	$14	Source10/02/2026
	Together	Via OpenRoutertogether	$2.7	$13.5	Source10/02/2026
	Wafer	Via OpenRouterwafer	$1.4417	$14	Source10/02/2026
	Wafer	Via OpenRouterwafer/us	$2.8	$14	Source10/02/2026

Performance

Shared benchmarks

Only published results with the same benchmark version, task, metric, and comparison cohort are paired. Different reasoning levels, agents, harnesses, or output limits appear directly in the table.

Capability profile from matched benchmarks

Each axis averages directly matched benchmark families. A value of 100 means the stronger value within this pair, not a universal quality score.

Gemini 3.1 Pro Preview

Kimi K3

Coding: 65.5 / 98 (7 benchmark families)

Reasoning: 100 / 97.4 (1 benchmark family)

Mathematics: 29.9 / 100 (1 benchmark family)

Knowledge: 100 / 96.7 (1 benchmark family)

Finance: 83.3 / 99.3 (6 benchmark families)

Legal: 48.9 / 99.5 (3 benchmark families)

Agentic tasks: 87.5 / 100 (1 benchmark family)

Multimodal: 100 / 99.9 (1 benchmark family)

30 matched rows from 28 benchmark families were considered. The radar shows 8 of 13 comparable categories.

Gemini 3.1 Pro PreviewKimi K3

BioMysteryBenchOverall. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview60 %

Kimi K371.481 %

Code MigrationOverall. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview17.313 %

Kimi K316.099 %

CorpFin v2Overall. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview64.491 %

Kimi K371.562 %

CyberBench v1.1Patch. Higher is better. Test setup not documented as identical

Gemini 3.1 Pro Preview69.492 %

Kimi K383.051 %

EMBOverall. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview52.616 %

Kimi K366.399 %

Finance Agent (v2)Overall. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview42.982 %

Kimi K354.36 %

GPQA DiamondOverall. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview95.454 %

Kimi K392.929 %

Harvey's Legal Agent BenchmarkOverall · Task fully resolved. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview0 %

Kimi K310.833 %

IOIOverall. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview51.833 %

Kimi K348.944 %

Legal Research BenchOverall · All-pass. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview20.673 %

Kimi K344.231 %

LegalBenchOverall. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview87.398 %

Kimi K386.016 %

LiveCodeBench v6Overall. Higher is better. No documented setup difference

Gemini 3.1 Pro Preview88.485 %

Kimi K387.185 %

Key to the marks:better valuebetter, but not conclusivetie

Benchmark scores for Gemini 3.1 Pro Preview and Kimi K3
Benchmark and task	Gemini 3.1 Pro Preview	Kimi K3	Source
BioMysteryBenchOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Harness: Terminus 2	60 %	Winner71.481 %	Vals AI Source details Data as of: 09/27/2026
Code MigrationOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner17.313 %	16.099 %	Vals AI Source details Data as of: 09/27/2026
CorpFin v2Overall Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better	64.491 %	Winner71.562 %	Vals AI Source details Data as of: 08/12/2026
CyberBench v1.1Patch Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterKimi K3: Temperature: 1, Output limit: 262,144 tokens	69.492 %	Higher measured value83.051 %	Vals AI Source details Data as of: 08/03/2026
EMBOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	52.616 %	Winner66.399 %	Vals AI Source details Data as of: 09/27/2026
Finance Agent (v2)Overall Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better	42.982 %	Winner54.36 %	Vals AI Source details Data as of: 09/27/2026
GPQA DiamondOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner95.454 %	92.929 %	Vals AI Source details Data as of: 09/01/2026
Harvey's Legal Agent BenchmarkOverall · Task fully resolved Test details No documented difference in the test setupVersion: 1Metric: task resolution rateScoring: Higher is better	0 %	Winner10.833 %	Vals AI Source details Data as of: 09/27/2026
IOIOverall Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better	Winner51.833 %	48.944 %	Vals AI Source details Data as of: 09/27/2026
Legal Research BenchOverall · All-pass Test details No documented difference in the test setupVersion: 1Metric: all-pass rateScoring: Higher is better	20.673 %	Winner44.231 %	Vals AI Source details Data as of: 09/27/2026
LegalBenchOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner87.398 %	86.016 %	Vals AI Source details Data as of: 09/27/2026
LiveCodeBench v6Overall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner88.485 %	87.185 %	Vals AI Source details Data as of: 09/01/2026
LMArena Text, Style ControlOverall Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGemini 3.1 Pro Preview: Snapshot: gemini-3.1-pro-preview, Harness: Arena text, style control Kimi K3: Snapshot: kimi-k3-max, Reasoning: max, Harness: Arena text, style controlGemini 3.1 Pro Preview: Style-controlled Arena rating from 119,196 anonymous pairwise votes.Kimi K3: Style-controlled Arena rating from 26,400 anonymous pairwise votes.	1,486.678 95% confidence interval: ±3.034The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	Higher measured value1,487.964 95% confidence interval: ±4.866The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/25/2026
MedCodeOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner59.062 %	48.884 %	Vals AI Source details Data as of: 09/26/2026
MedScribeOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	76.114 %	Winner87.958 %	Vals AI Source details Data as of: 09/26/2026
MMLU-ProOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner90.987 %	87.973 %	Vals AI Source details Data as of: 09/01/2026
MMMU-ProOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner88.208 %	88.15 %	Vals AI Source details Data as of: 09/01/2026
MortgageTaxOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner69.396 %	66.336 %	Vals AI Source details Data as of: 09/01/2026
ProgramBenchOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	0 %	Winner2 %	Vals AI Source details Data as of: 09/27/2026
ProofBenchOverall Test details No documented difference in the test setupVersion: 1.1Metric: accuracyScoring: Higher is better	26 %	Winner87 %	Vals AI Source details Data as of: 09/28/2026
Public Benefits Bench v1.1Overall Test details No documented difference in the test setupVersion: 1.1Metric: accuracyScoring: Higher is better	53.789 %	Winner68.268 %	Vals AI Source details Data as of: 09/27/2026
SAGEOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	48.677 %	Winner54.26 %	Vals AI Source details Data as of: 09/26/2026
SWE-bench VerifiedOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	78.8 %	Winner93.4 %	Vals AI Source details Data as of: 09/01/2026
Tax Agent BenchOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	37.842 %	Winner68.672 %	Vals AI Source details Data as of: 09/27/2026
TaxEval v2Overall Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better	72.882 %	Winner75.716 %	Vals AI Source details Data as of: 09/01/2026
Terminal-Bench 2.1Overall Test details No documented difference in the test setupVersion: 2.1Metric: accuracyScoring: Higher is better	70.787 %	Winner80.899 %	Vals AI Source details Data as of: 09/27/2026
Vals IndexOverall Test details No documented difference in the test setupVersion: 2Metric: weighted index scoreScoring: Higher is better	41.903 %	Winner57.813 %	Vals AI Source details Data as of: 09/27/2026
Vals Multimodal IndexOverall Test details No documented difference in the test setupVersion: 1.2Metric: weighted index scoreScoring: Higher is better	56.068 %	Winner73.42 %	Vals AI Source details Data as of: 08/11/2026
Vibe Code Bench 1-100Overall Test details No documented difference in the test setupVersion: 1.0Metric: accuracyScoring: Higher is betterSettings: Harness: OpenHands	6.694 %	Winner18.244 %	Vals AI Source details Data as of: 09/22/2026
Vibe Code Bench v1.1Overall Test details No documented difference in the test setupVersion: 1.1Metric: accuracyScoring: Higher is betterSettings: Harness: OpenHands	32.034 %	Winner84.963 %	Vals AI Source details Data as of: 09/27/2026
CorpFin v2Exact Pages Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	60.839 %	Higher measured value69.347 %	Vals AI Source details Data as of: 08/06/2026
CyberBench v1.1PoC Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterKimi K3: Temperature: 1, Output limit: 262,144 tokens	3.39 %	Higher measured value75 %	Vals AI Source details Data as of: 08/03/2026
EMBComps Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	44.505 %	Higher measured value72.685 %	Vals AI Source details Data as of: 08/06/2026
Finance Agent (v2)Adjustments Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	38.361 %	Higher measured value51.461 %	Vals AI Source details Data as of: 08/06/2026
GPQA DiamondFew-Shot CoT Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,535 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	Higher measured value95.96 %	93.434 %	Vals AI Source details Data as of: 08/09/2026
Legal Research BenchAdministrative / Regulatory · All-pass Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: all-pass rateScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	24.242 %	Higher measured value54.545 %	Vals AI Source details Data as of: 08/06/2026
LegalBenchConclusion Tasks Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	89.257 %	Higher measured value90.733 %	Vals AI Source details Data as of: 08/09/2026
LiveCodeBench v6Easy Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	Higher measured value99.068 %	98.758 %	Vals AI Source details Data as of: 08/09/2026
LMArena Text, Style ControlBusiness, management and finance Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGemini 3.1 Pro Preview: Snapshot: gemini-3.1-pro-preview, Harness: Arena text, style control Kimi K3: Snapshot: kimi-k3-max, Reasoning: max, Harness: Arena text, style controlGemini 3.1 Pro Preview: Style-controlled Arena rating from 23,398 anonymous pairwise votes.Kimi K3: Style-controlled Arena rating from 5,149 anonymous pairwise votes.	1,477.195 95% confidence interval: ±5.145The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	Higher measured value1,489.173 95% confidence interval: ±8.766The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/25/2026
MMLU-ProBiology Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,535 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	Higher measured value95.258 %	93.863 %	Vals AI Source details Data as of: 08/09/2026
MortgageTaxNumerical Extraction Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	76.391 %	Higher measured value78.855 %	Vals AI Source details Data as of: 08/09/2026
ProgramBenchAlmost Resolved Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	1 %	Higher measured value17.5 %	Vals AI Source details Data as of: 08/09/2026
SAGEAutomata Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 30,000 tokens Kimi K3: Temperature: 1, Output limit: 30,000 tokens	32.143 %	Higher measured value44.048 %	Vals AI Source details Data as of: 08/09/2026
SWE-bench Verified<15 min fix Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	88.66 %	Higher measured value92.268 %	Vals AI Source details Data as of: 08/08/2026
TaxEval v2Correctness Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,535 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	60.749 %	Higher measured value65.331 %	Vals AI Source details Data as of: 08/09/2026
Terminal-Bench 2.1Easy Test details Test setup not documented as identicalNo direct winnerVersion: 2.1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	No direct winner91.667 %	No direct winner91.667 %	Vals AI Source details Data as of: 08/06/2026
Vals IndexCorpFin v2 Test details Test setup not documented as identicalNo direct winnerVersion: 1.2Metric: benchmark scoreScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	65.734 %	Higher measured value72.611 %	Vals AI Source details Data as of: 08/09/2026
Vals Multimodal IndexCorpFin v2 Test details Test setup not documented as identicalNo direct winnerVersion: 1.2Metric: benchmark scoreScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	65.734 %	Higher measured value72.611 %	Vals AI Source details Data as of: 08/09/2026
CorpFin v2Max Fitting Context Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	66.9 %	Higher measured value72.727 %	Vals AI Source details Data as of: 08/06/2026
EMBDataroom Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	79.17 %	Higher measured value85.13 %	Vals AI Source details Data as of: 08/06/2026
Finance Agent (v2)All-Pass Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	30.879 %	Higher measured value41.69 %	Vals AI Source details Data as of: 08/06/2026
GPQA DiamondZero-Shot CoT Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,535 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	Higher measured value94.949 %	92.424 %	Vals AI Source details Data as of: 08/09/2026
Legal Research BenchAdministrative / Regulatory · Weighted partial credit Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: weighted partial-credit scoreScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	79.215 %	Higher measured value90.731 %	Vals AI Source details Data as of: 08/06/2026
LegalBenchInterpretation Tasks Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	Higher measured value83.823 %	82.612 %	Vals AI Source details Data as of: 08/09/2026
LiveCodeBench v6Hard Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	Higher measured value76.571 %	74.286 %	Vals AI Source details Data as of: 08/09/2026
LMArena Text, Style ControlChinese Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGemini 3.1 Pro Preview: Snapshot: gemini-3.1-pro-preview, Harness: Arena text, style control Kimi K3: Snapshot: kimi-k3-max, Reasoning: max, Harness: Arena text, style controlGemini 3.1 Pro Preview: Style-controlled Arena rating from 8,183 anonymous pairwise votes.Kimi K3: Style-controlled Arena rating from 1,940 anonymous pairwise votes.	1,528.226 95% confidence interval: ±7.531The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	Higher measured value1,539.78 95% confidence interval: ±13.819The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/25/2026
MMLU-ProBusiness Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,535 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	Higher measured value93.409 %	92.015 %	Vals AI Source details Data as of: 08/09/2026
MortgageTaxSemantic Extraction Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	Higher measured value62.401 %	53.816 %	Vals AI Source details Data as of: 08/09/2026
ProgramBenchFully Resolved Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	0 %	Higher measured value2 %	Vals AI Source details Data as of: 08/09/2026
SAGECalculus Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 30,000 tokens Kimi K3: Temperature: 1, Output limit: 30,000 tokens	Higher measured value78.395 %	75.617 %	Vals AI Source details Data as of: 08/09/2026
SWE-bench Verified>4 hours Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	33.333 %	Higher measured value66.667 %	Vals AI Source details Data as of: 08/08/2026
TaxEval v2Stepwise Reasoning Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,535 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	85.016 %	Higher measured value86.1 %	Vals AI Source details Data as of: 08/09/2026
Terminal-Bench 2.1Hard Test details Test setup not documented as identicalNo direct winnerVersion: 2.1Metric: accuracyScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	60 %	Higher measured value74.444 %	Vals AI Source details Data as of: 08/06/2026
Vals IndexFinance Agent v2 Test details Test setup not documented as identicalNo direct winnerVersion: 1.2Metric: benchmark scoreScoring: Higher is betterGemini 3.1 Pro Preview: Reasoning: high, Temperature: 1, Output limit: 65,536 tokens Kimi K3: Temperature: 1, Output limit: 262,144 tokens	42.968 %	Higher measured value55.882 %	Vals AI Source details Data as of: 08/09/2026
Key to the marks:better valuebetter, but not conclusivetie

Our model tests

Our tests

Compare how the models respond to the same prompt. Each test shows the first attempt, with no subsequent fixes to the generated code. These results do not contribute to an overall score.

Pelican on a bicycle

A community classic for free-form SVG drawing.

Gemini 3.1 Pro Preview

API access was unavailable for this run. The model was not evaluated.

Reasoning (requested): high

First attempt

Run date: 7 Sept 2026
API provider: openrouter
Requested API model: google/gemini-3.1-pro-preview
Reasoning setting: high
Pinned endpoint: Google AI Studio | google/gemini-3.1-pro-preview-20260219
Temperature: Provider default
Token limit including reasoning: 8,192 tokens
Test protocol: community-visual-tests-v1
Technical checks: Not run

Not yet visually reviewed. A successful recording does not confirm correct physics or full compliance with the prompt.

Kimi K3

API access was unavailable for this run. The model was not evaluated.

Reasoning (requested): high

First attempt

Run date: 7 Sept 2026
API provider: openrouter
Requested API model: moonshotai/kimi-k3
Reasoning setting: high
Pinned endpoint: Moonshot AI | moonshotai/kimi-k3-20260715
Temperature: Provider default
Token limit including reasoning: 8,192 tokens
Test protocol: community-visual-tests-v1
Technical checks: Not run

Not yet visually reviewed. A successful recording does not confirm correct physics or full compliance with the prompt.

Prompt and test conditions

Generate an SVG of a pelican riding a bicycle

Token limit including reasoning: 8,192 tokens

Task origin (Simon Willison)

How to read this comparison

Winning one benchmark is not an overall verdict

Your workload, budget, and required context length matter most. A coding benchmark says little about visual understanding or agent performance.

The standard price table uses direct developer rates. The provider comparison labels OpenRouter endpoints, regions, and special tariffs separately.

Every result links to its measurement source and retrieval date. Labels distinguish vendor reports, official benchmark leaderboards, and independent evaluations. Disputed or archived results are excluded.

More matchups

Compare Gemini 3.1 Pro Preview and Kimi K3 with other leading models using the same data and benchmark logic.

View the complete LLM comparison