Moonshot AI: Kimi K3

Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work

Source-linked kimi-k3 Open weights Released 2026-07-16

API record Report

InputT

OutputT

Input price$3/M

Output price$15/M

Context1.04858M

Max output131.072K

Providers86

Inference availability

Providers

Provider-specific identifiers, limits, and listed prices per million tokens. Every row links back to the provider's own documentation.

Report a price

Providers offering Kimi K3
Provider	Provider model ID	Context	Max output	Input	Output	Cache read	Capabilities	Docs
Kimi For Coding	`k3`	1.04858M	131.072K	—	—	—	ReasoningToolsJSON	Docs ↗
Umans AI Coding Plan	`umans-kimi-k3`	1.04858M	131.072K	—	—	—	ReasoningToolsJSON	Docs ↗
SCNet Token Plan	`Kimi-K3`	1.04858M	131.072K	—	—	—	ReasoningToolsJSON	Docs ↗
Nvidia	`moonshotai/kimi-k3`	1.04858M	131.072K	—	—	—	ReasoningToolsJSON	Docs ↗
SenseNova (China)	`kimi-k3`	1.04858M	65.536K	—	—	—	ReasoningToolsJSON	Docs ↗
Volcengine Ark Coding Plan	`kimi-k3`	1.04858M	131.072K	—	—	—	ReasoningToolsJSON	Docs ↗
Kimi For Coding (kimi.ai)	`k3`	1.04858M	131.072K	—	—	—	ReasoningToolsJSON	Docs ↗
Kimi For Coding (kimi.com)	`k3`	1.04858M	131.072K	—	—	—	ReasoningToolsJSON	Docs ↗
OpenRouter	`moonshotai/kimi-k3`	1.04858M	943.718K	$0.239	$10.8	$0.21	ReasoningToolsJSON	Docs ↗
Kenari	`kimi-k3`	1.04858M	131.072K	$0.857	$8.571	$0.114	ReasoningToolsJSON	Docs ↗
NanoGPT	`moonshotai/kimi-k3`	1.04858M	943.718K	$1.2	$5.5	$0.2	ReasoningToolsJSON	Docs ↗
engy	`kimi-k3`	1.11309M	65.536K	$1.95	$9.75	$0.195	ReasoningToolsJSON	Docs ↗
CrofAI	`kimi-k3`	1M	262.144K	$2	$8	$0.25	ReasoningToolsJSON	Docs ↗
ainetcafe	`Kimi-K3`	262.144K	32.768K	$2.1	$10.5	$0.3	ReasoningToolsJSON	Docs ↗
Boundless	`kimi-k3`	1.04858M	131.072K	$2.3	$11.4	$0.23	ReasoningToolsJSON	Docs ↗
IteraCompute	`moonshotai/kimi-k3`	1.04858M	999.999K	$2.35	$11.76	$0.23	ReasoningToolsJSON	Docs ↗
Vancine	`kimi-k3`	1.04858M	131.072K	$2.4	$12	$0.24	ReasoningToolsJSON	Docs ↗
Wallaby	`moonshotai/kimi-k3`	1.04858M	131.072K	$2.7	$13.5	$0.27	ReasoningToolsJSON	Docs ↗
SiliconFlow	`moonshotai/Kimi-K3`	1.04858M	262K	$2.7	$13.5	$0.27	ReasoningToolsJSON	Docs ↗
Alibaba (China)	`kimi-k3`	1.04858M	1.04858M	$2.827	$14.133	$0.283	ReasoningToolsJSON	Docs ↗
Deep Infra	`moonshotai/Kimi-K3`	1.04858M	131.072K	$2.85	$14.25	$0.285	ReasoningToolsJSON	Docs ↗
Moonshot AI (China)	`kimi-k3`	1.04858M	1.04858M	$2.988	$14.94	$0.299	ReasoningToolsJSON	Docs ↗
DevPass (LLM Gateway)	`kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
OpenCode Go	`kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Moonshot AI	`kimi-k3`	1.04858M	1.04858M	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Vercel AI Gateway	`moonshotai/kimi-k3`	1M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
EmpirioLabs AI	`kimi-k3`	1M	131.072K	$3	$15	$3	ReasoningToolsJSON	Docs ↗
Cloudflare AI Gateway	`moonshotai/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
ZenMux	`moonshotai/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
CrossModel	`moonshot/kimi-k3`	1.04858M	1.04858M	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Novita AI	`moonshotai/kimi-k3`	1.04858M	1.04858M	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
ClinePass	`cline-pass/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Fireworks AI	`accounts/fireworks/models/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
CoreWeave	`moonshotai/Kimi-K3`	1.04858M	1.04858M	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Baseten	`moonshotai/Kimi-K3`	1.04858M	262.144K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Nebius Token Factory	`moonshotai/Kimi-K3`	1.024M	8K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
OpenCode	`kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
ai&	`moonshotai/kimi-k3`	1.04858M	131.072K	$3	$12.5	$0.5	ReasoningToolsJSON	Docs ↗
Together AI	`moonshotai/Kimi-K3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
AIHubMix	`kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Synthetic	`hf:moonshotai/Kimi-K3`	524.288K	65.536K	$3	$15	$0.45	ReasoningToolsJSON	Docs ↗
Hugging Face	`moonshotai/Kimi-K3`	1M	131.072K	$3	$15	—	ReasoningToolsJSON	Docs ↗
Cortecs	`kimi-k3`	1.04858M	1.04858M	$3	$14.999	—	ReasoningToolsJSON	Docs ↗
Perplexity Agent	`moonshot-ai/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Modal	`moonshotai/Kimi-K3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Merge Gateway	`moonshot/kimi-k3`	1.04858M	1.04858M	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Impossibl	`moonshotai/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
DigitalOcean	`kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Kilo Gateway	`moonshotai/kimi-k3`	1.04858M	943.718K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
TensorX	`moonshotai/kimi-k3`	1.04858M	131.072K	$3	$15	$0.75	ReasoningToolsJSON	Docs ↗
Umans AI	`umans-kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Neuralwatt	`kimi-k3`	1.04856M	1.04856M	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Requesty	`kimi-k3`	1.04858M	262.144K	$3	$15	$0.45	ReasoningToolsJSON	Docs ↗
Pioneer	`moonshotai/Kimi-K3`	1M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
GitHub Copilot	`kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Ofox	`moonshotai/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Vivgrid	`kimi-k3`	1M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Neon	`kimi-k3`	1.04858M	65.536K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
CoralBricks	`kimi-k3`	1.04858M	131.072K	$3	$15	—	ReasoningToolsJSON	Docs ↗
Eden AI	`moonshot/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Abacus	`moonshotai/Kimi-K3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Jalapeno Cloud	`Kimi-K3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
TokenGo	`moonshotai/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
OrcaRouter	`kimi/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Opper	`moonshot/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Arcee	`moonshotai/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
302.AI	`kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Alibaba	`kimi-k3`	1.04858M	1.04858M	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Opper	`kimi-k3`	1.04858M	131.072K	$3	$15	—	ReasoningToolsJSON	Docs ↗
Tempr Gateway	`moonshot/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Moark	`kimi-k3`	1.04858M	131.072K	$3	$15	—	ReasoningToolsJSON	Docs ↗
AI/ML API	`moonshot/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
CommandCode	`moonshotai/Kimi-K3`	1M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Standard Compute	`moonshotai/kimi-k3`	1.04858M	50K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Perplexity Agent	`perplexity/kimi-k3`	1.04858M	131.072K	$3	$15	$0.3	ReasoningToolsJSON	Docs ↗
Melious	`kimi-k3`	1M	131.072K	$3.196	$15.981	$0.841	ReasoningToolsJSON	Docs ↗
Charm Hyper	`kimi-k3`	1.04858M	16K	$3.266	$16.332	$0.327	ReasoningToolsJSON	Docs ↗
NEAR AI Cloud	`moonshotai/kimi-k3`	1.04858M	131.072K	$3.3	$16.5	$0.33	ReasoningToolsJSON	Docs ↗
Berget.AI	`moonshotai/Kimi-K3`	327.68K	32.768K	$3.365	$16.823	$1.009	ReasoningToolsJSON	Docs ↗
Prime Inference	`moonshotai/kimi-k3`	1.04858M	943.718K	$3.45	$17.25	—	ReasoningToolsJSON	Docs ↗
Venice AI	`kimi-k3`	1M	131.072K	$3.75	$18.75	$0.375	ReasoningToolsJSON	Docs ↗
GreenPT	`kimi-k3`	1.04858M	131.072K	$3.762	$18.81	$0.941	ReasoningToolsJSON	Docs ↗
Tinfoil	`kimi-k3`	262.144K	131.072K	$4	$20	$0.8	ReasoningToolsJSON	Docs ↗
DevPass (LLM Gateway)	`kimi-k3-fast`	1.04038M	131.072K	$4.5	$22.5	$0.45	ReasoningToolsJSON	Docs ↗
Pioneer	`moonshotai/Kimi-K3-Fast`	1M	131.072K	$4.5	$22.5	$0.45	ReasoningToolsJSON	Docs ↗
Inco	`kimi-k3:fast`	1.04858M	131.072K	$6	$30	$0.6	ReasoningToolsJSON	Docs ↗

Capability badges appear only where the provider catalog explicitly lists support. A blank cell means the source is silent, not that the feature is absent.

Listed rates

Price across providers

Input price per million tokens as published by each provider. Bars are drawn from listed rates only — no traffic weighting, since the catalog observes no requests.

Lowest input $0.239/M

Across 78 priced providers

Median input $3/M

Midpoint of listed rates

Highest input $6/M

25.1× the lowest listed rate

Output range $5.5 – $30

Per million output tokens

OpenRouter $0.239/MLowest

Kenari $0.857/M

NanoGPT $1.2/M

engy $1.95/M

CrofAI $2/M

ainetcafe $2.1/M

Boundless $2.3/M

IteraCompute $2.35/M

Vancine $2.4/M

Wallaby $2.7/M

SiliconFlow $2.7/M

Alibaba (China) $2.827/M

Deep Infra $2.85/M

Moonshot AI (China) $2.988/M

DevPass (LLM Gateway) $3/M

OpenCode Go $3/M

Moonshot AI $3/M

Vercel AI Gateway $3/M

EmpirioLabs AI $3/M

Cloudflare AI Gateway $3/M

ZenMux $3/M

CrossModel $3/M

Novita AI $3/M

ClinePass $3/M

Fireworks AI $3/M

CoreWeave $3/M

Baseten $3/M

Nebius Token Factory $3/M

OpenCode $3/M

ai& $3/M

Together AI $3/M

AIHubMix $3/M

Synthetic $3/M

Hugging Face $3/M

Cortecs $3/M

Perplexity Agent $3/M

Modal $3/M

Merge Gateway $3/M

Impossibl $3/M

DigitalOcean $3/M

Kilo Gateway $3/M

TensorX $3/M

Umans AI $3/M

Neuralwatt $3/M

Requesty $3/M

Pioneer $3/M

GitHub Copilot $3/M

Ofox $3/M

Vivgrid $3/M

Neon $3/M

CoralBricks $3/M

Eden AI $3/M

Abacus $3/M

Jalapeno Cloud $3/M

TokenGo $3/M

OrcaRouter $3/M

Opper $3/M

Arcee $3/M

302.AI $3/M

Alibaba $3/M

Opper $3/M

Tempr Gateway $3/M

Moark $3/M

AI/ML API $3/M

CommandCode $3/M

Standard Compute $3/M

Perplexity Agent $3/M

Melious $3.196/M

Charm Hyper $3.266/M

NEAR AI Cloud $3.3/M

Berget.AI $3.365/M

Prime Inference $3.45/M

Venice AI $3.75/M

GreenPT $3.762/M

Tinfoil $4/M

DevPass (LLM Gateway) $4.5/M

Pioneer $4.5/M

Inco $6/M

Cost calculator

Estimate a workload

Provider Input tokens 1,000,000 Output tokens 250,000 Cached input 0%

$0.00

Excludes taxes, non-token charges, and tiered discounts.

Context limits also differ by provider, from 262.144K to 1.11309M tokens. Compare the provider table above before choosing on price alone.

Specification

Capabilities

Recorded from the source catalog and provider listings.

✓ Reasoning Yes

✓ Tool calling Yes

✓ Structured output Yes

✓ Attachments Yes

✓ Vision input Yes

✓ Open weights Yes

Creator: Moonshot AI
Model family: kimi-k3
Knowledge cutoff: Not documented
License: Not documented
Release date: 2026-07-16
Model ID: moonshotai/kimi-k3

Published evaluations

Benchmarks

Every result stays attached to its source, version, metric and harness. Scores from different versions are never merged, and the profile below plots each benchmark against its own population rather than on a shared scale.

Benchmark registry

Benchmark profile

Each axis is this model's percentile among the 8 benchmarks it has published results for, measured against every other model with a score on that same benchmark. Percentiles are used because benchmarks do not share a scale — a 60 on one is not a 60 on another. Hover any point for the raw score.

Intelligence Index index Source ↗

Intelligence Index #31 of 334

43.6

Coding Index index Source ↗

Coding Index #16 of 214

76.2

Agentic Index index Source ↗

Agentic Index #20 of 201

50.0

Design Arena: website elo Source ↗

Design Arena: website #3 of 200

1334.0

Design Arena: gamedev elo Source ↗

Design Arena: gamedev #13 of 191

1359.0

Design Arena: codecategories elo Source ↗

Design Arena: codecategories #4 of 188

1359.0

Design Arena: dataviz elo Source ↗

Design Arena: dataviz #6 of 183

1351.0

Design Arena: uicomponent elo Source ↗

Design Arena: uicomponent #10 of 182

1339.0

Design Arena: 3d elo Source ↗

Design Arena: 3d #11 of 176

1389.0

Design Arena: svg elo Source ↗

Design Arena: svg #8 of 128

1302.0

Design Arena: webapps elo Source ↗

Design Arena: webapps #11 of 80

1295.0

Design Arena: mobileapps elo Source ↗

Design Arena: mobileapps #7 of 77

1270.0

Design Arena: fullstack elo Source ↗

Design Arena: fullstack #6 of 69

1315.0

Design Arena: androidnative elo Source ↗

Design Arena: androidnative #5 of 61

1256.0

Design Arena: godotgamedev elo Source ↗

Design Arena: godotgamedev #24 of 55

1223.0

Design Arena: python-pptxslides elo Source ↗

Design Arena: python-pptxslides #15 of 48

1219.0

Design Arena: agenticgamedev elo Source ↗

Design Arena: agenticgamedev #5 of 44

1252.0

Design Arena: htmlslides elo Source ↗

Design Arena: htmlslides #15 of 43

1153.0

Design Arena: blender3d elo Source ↗

Design Arena: blender3d #16 of 35

1209.0

Design Arena: spreadsheet elo Source ↗

Design Arena: spreadsheet #17 of 31

1115.0

BrowseComp accuracy · 2026-07-16 Source ↗

BrowseComp #2 of 15

91.2

JobBench score · 2026-07-16 Source ↗

JobBench #6 of 8

52.9

CharXiv Reasoning accuracy · 2026-07-16 Source ↗

CharXiv Reasoning #1 of 7

91.3

AutomationBench success rate · 2026-07-16 Source ↗

AutomationBench #2 of 5

30.8

GDPval-AA — v2 Elo · 2026-07-16 Source ↗

GDPval-AA #2 of 4

1668.0

ZeroBench pass@5 · 2026-07-16 Source ↗

ZeroBench #2 of 3

41.0

AA-Briefcase Elo · 2026-07-16 Source ↗

1548.0 2 models scored — too few for a distribution

DeepSWE — 1.1 resolve rate · Kimi Code · 2026-07-16 Source ↗

67.5 1 model scored — too few for a distribution

FrontierSWE dominance score · Kimi Code · 2026-07-16 Source ↗

81.2 1 model scored — too few for a distribution

Program Bench score · Kimi Code · 2026-07-16 Source ↗

77.8 1 model scored — too few for a distribution

SWE Marathon — 1.1 resolve rate · Claude Code · 2026-07-16 Source ↗

42.0 1 model scored — too few for a distribution

SpreadsheetBench — 2 score · Claude Code · 2026-07-16 Source ↗

34.8 1 model scored — too few for a distribution

Terminal-Bench — 2.1 accuracy · Kimi Code · 2026-07-16 Source ↗

88.3 1 model scored — too few for a distribution

Each strip shows every published score for that benchmark, with this model marked. The lighter shape behind the ticks is the density of results, and the dashed line is the median.

Cost against capability

Price and performance

Listed input price plotted against Intelligence Index, the benchmark with the widest published coverage that this model appears in.

The stepped line is the efficient frontier: at each price, the best score available for that money or less. A model sitting on it is not being beaten by anything cheaper. Price is log-scaled because listed rates span four orders of magnitude. Only models with both a listed price and a score on this benchmark can appear.

Catalog activity

Change log

Field-level changes detected between successful source imports.

Full change log

Oct 24, 2026Price Completion10.8 → 15.0

Oct 24, 2026Max Output Tokens943718 → 131072

Oct 24, 2026Price Prompt0.23900000000000002 → 3.0

Oct 24, 2026Max Output Tokens131072 → 943718

Oct 24, 2026Price Completion15.0 → 10.8

Oct 24, 2026Price Prompt3.0 → 0.23900000000000002

Oct 24, 2026Max Output Tokens943718 → 131072

Oct 24, 2026Price Completion10.8 → 15.0

Oct 24, 2026Price Prompt0.23900000000000002 → 3.0

Oct 24, 2026Max Output Tokens131072 → 943718

Provenance

Sources & verification

Every figure on this page traces back to one of these records.

Methodology

Last verified: Oct 24, 2026
Status: Source-linked
Confidence: Medium
Catalog source: Models.dev

Public API

Use this record

Fetch the complete source-linked model record. No key, no account, no rate-limited tier.

API documentation

Endpoint

GET https://modelbench.lol/api/v1/models/moonshotai/kimi-k3

curl

curl "https://modelbench.lol/api/v1/models/moonshotai/kimi-k3"

Common questions

Frequently asked questions

Answered directly from the stored record — nothing here is generated beyond the catalog's own fields.

What is Kimi K3?

Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work. It is published by Moonshot AI and catalogued here from Models.dev.

How much does Kimi K3 cost?

Listed input pricing starts at $0.239 per million tokens from OpenRouter, rising to $6 across 78 listed providers.

What is the context length of Kimi K3?

Kimi K3 accepts up to 1.04858M tokens of context and returns up to 131.072K output tokens.

Does Kimi K3 support tool calling and structured output?

Provider catalogs list support for tool calling, structured output, reasoning, and image input.

Which providers serve Kimi K3?

82 providers list this model: Kimi For Coding, Umans AI Coding Plan, SCNet Token Plan, Nvidia, SenseNova (China), Volcengine Ark Coding Plan and 76 more.

Are the weights for Kimi K3 open?

Yes. The weights are published.

When was Kimi K3 released?

The catalog records a release date of 2026-07-16, last verified Oct 24, 2026.