Model evidence profile

Kimi K3

Name: Kimi K3 benchmark evidence
Creator: Moonshot AI
Published: 2026-08-13T07:56:45.000Z

Moonshot AI · Unknown · Reasoning

currentsupported

Overall public score: 80.21
Source rank: #5
Evidence coverage: 5 benchmarks · 2 sources

Strongest published evidencePublic overall score 80.21 at source rank #5.

Validate before choosingValidate the selected route price, context limits, and evidence freshness before choosing.

Revision benchmark_be54b952bd03dda253cf3c67b5ae12f1 · Published Aug 13, 2026 · Checked Aug 13, 2026 · stale

Relative field position

Capability radar

Percentiles use eligible source ranks. Missing axes remain unavailable and never become zero.

Agentic percentile: 96.9 percentileRank #5 of 128
Coding percentile: 97.0 percentileRank #5 of 133
Knowledge percentile: 91.7 percentileRank #5 of 49
Multimodal grounded percentile: 93.8 percentileRank #3 of 33
Overall percentile: Overall: Unavailable
Reasoning percentile: Reasoning: Unavailable

Published measurements

Category scores

Scores retain their source units; percentile and rank appear only for eligible fields.

Agenticsupported

73.7

Rank: #5 of 128
Percentile: 96.9%
Benchmarks: 5

Codingsupported

77.5

Rank: #5 of 133
Percentile: 97.0%
Benchmarks: 5

Knowledgesupported

84.8

Rank: #5 of 49
Percentile: 91.7%
Benchmarks: 5

Multimodal groundedsupported

87.9

Rank: #3 of 33
Percentile: 93.8%
Benchmarks: 5

Overallsupported

80.2

Rank: #5
Percentile: Unavailable
Benchmarks: 5

Route-specific facts

Pricing and specifications

Conflicting routes remain separate and attributable.

moonshot-aibenchlm:kimi-3

primary

Input / 1M: $3.00
Cached input / 1M: $0.30
Output / 1M: $15.00
Context: 1,050,000

View price source

Context window: 1,050,000
Maximum output: Unavailable
Input modalities: Unavailable
Output modalities: Unavailable
Release date: 2026-07-16
Self hosting: Not verified

Auditable evidence

Benchmark ledger

Display values, source ranks, and provenance remain visible without implying unsupported aggregate weight.

agentic

Score	Rank	Weight	Last Updated	Source
73.67	#5	Not published	Aug 13, 2026	BenchLM

coding

Score	Rank	Weight	Last Updated	Source
77.52	#5	Not published	Aug 13, 2026	BenchLM

knowledge

Score	Rank	Weight	Last Updated	Source
84.80	#5	Not published	Aug 13, 2026	BenchLM

multimodalGrounded

Score	Rank	Weight	Last Updated	Source
87.90	#3	Not published	Aug 13, 2026	BenchLM

overall

Score	Rank	Weight	Last Updated	Source
80.21	#5	Not published	Aug 13, 2026	BenchLM