跳轉到主要內容

Kimi K3 (max)

KimiKimi開源權重Kimi K3 License · 商用許可

描述

Kimi K3 is Moonshot AI's flagship open Mixture-of-Experts model for long-horizon coding, knowledge work, and reasoning. It has 2.8 trillion total parameters, activates 16 of 896 experts, and combines Kimi Delta Attention, Attention Residuals, and Stable LatentMoE. The model supports a 1M-token context window, native image and video understanding, tool calling, structured output, and always-on reasoning with max thinking effort at launch.

發布日期
2026-07-16
參數規模
2.8T
上下文長度
1.0M
支援模態
image, text, video

能力雷達圖

57
general
74
coding
93
reasoning
72
science
70
agents
80
multimodal

排行榜排名

領域#排名分數來源
智慧體能力模型榜9
63.0
LS
程式碼能力榜2
98.0
AA
通用能力榜6
95.0
AA
多模態榜7
70.0
LS
科學能力8
93.0
AA

基準測試分數 (LLM Stats)

(LLM Stats (zeroeval))

Agents

AA-Briefcase1548.00 / 3000
Program Bench77.8%自報
DECK-Bench73.5%自報
Toolathlon73.2%自報
Kimi Code Bench v272.9%自報
OfficeQA Pro63.3%自報
Job Bench52.9%自報
MLS-Bench Lite48.3%自報
PostTrainBench36.6%自報
SpreadsheetBench 234.8%自報
AutomationBench30.8%自報

Code

FrontierSWE81.2%自報
DeepSWE 1.169.0%
DeepSWE67.5%自報
SWE-Marathon42.0%自報

Math

MathVision97.8%自報

Reasoning

GPQANYU + Cohere + Anthropic (2023)93.5%自報
CharXiv-R91.3%自報
BrowseCompOpenAI (2025)91.2%自報
Terminal-Bench 2.188.3%自報
MCP Atlas84.2%自報
Humanity's Last Exam56.0%自報
APEX-Agents37.6%自報

Search

DeepSearchQA95.0%自報

Vision

OmniDocBench91.1%自報
BabyVision85.7%自報
MMMU-Pro (with tools)83.4%自報
MMMU-Pro81.6%自報
PerceptionBench58.5%自報
WorldVQA51.0%自報
ZEROBench0.41 / 100自報

AA 評測指數

(Artificial Analysis)
Gpqa(NYU + Cohere + Anthropic (2023))
93.5
Terminalbench V2 1
85.0
Lcr(Artificial Analysis)
82.7
Coding Index(Artificial Analysis)
76.2
Intelligence Index(Artificial Analysis)
59.7
Scicode(UIUC + Argonne National Lab (2024))
58.7
Hle(Center for AI Safety + Scale AI (2025))
46.9
Tau Banking
46.0

LLM Stats 分類評分

(LLM Stats (zeroeval))
Productivity
100
Agents
92
Reasoning
78
General
60
Physics
90
Search
90
Biology
90
Chemistry
90
Math
80
Multimodal
80
Code
70
Tool Calling
70
Vision
70
Systems
40

定價

輸入價格$3 / 1M tokens
輸出價格$15 / 1M tokens
混合價格(3:1)$6 / 1M tokens
快取讀取價格$0.3 / 1M tokens

速度

Tokens/秒36.4
首Token延遲3.56s
首回答延遲58.53s

供應商價格排行

供應商價格排行

25 個供應商

最便宜: Moonshot AI最貴: Tinfoil
供應商輸入輸出
1Moonshot AI最便宜
$0
$0.00002
2Together
$0
$0.00002
3Novita
$0
$0.00002
4Fireworks
$0
$0.00002
5CrofAI
$2
$8
6Requesty
$2.25
$11.25
7Vancine
$2.4
$12
8DevPass (LLM Gateway)
$2.83
$14.13
9DigitalOcean
$2.85
$14.25
10Kimi主要
$3
$15
11OpenCode Go
$3
$15
12Vivgrid
$3
$15
13GitHub Copilot
$3
$15
14OpenCode Zen
$3
$15
15AIHubMix
$3
$15
16Moonshot AI (China)
$3
$15
17Cortecs
$3
$14.999
18Neuralwatt
$3
$15
19Neon
$3
$15
20EmpirioLabs AI
$3
$15
21CoralBricks
$3
$15
22Charm Hyper
$3.2664
$16.332
23Venice AI
$3.75
$18.75
24GreenPT
$3.762
$18.81
25Tinfoil
$4
$20

比較該模型在不同 API 供應商之間的定價。

外部連結