跳轉到主要內容

Kimi K2.6 (Non-reasoning)

KimiKimi開源權重Modified MIT License

描述

Kimi K2.6 is Moonshot AI's open-source, native multimodal agentic model focused on state-of-the-art coding, long-horizon execution, and agent swarm capabilities. It scales horizontally to 300 sub-agents executing 4,000 coordinated steps, dynamically decomposing tasks into parallel, domain-specialized subtasks. K2.6 unifies text, image, and video input with thinking and non-thinking modes, supports a 256K context, and powers proactive 24/7 background agents that manage schedules, execute code, and orchestrate cross-platform operations without human oversight.

發布日期
2026-04-20
參數規模
1.0T
上下文長度
262K
支援模態
image, text, video

能力雷達圖

70
general
60
coding
80
reasoning
60
science估算
60
agents
80
multimodal

缺少專門科學評測時,Science 由 LLM Stats 科學得分或推理能力估算。

排行榜排名

領域#排名分數來源
智慧體能力模型榜53
51.0
LS
數學推理46
84.0
LB
多模態榜59
55.0
LS

基準測試分數 (LLM Stats)

(LLM Stats (zeroeval))

Agents

MCP-Mark55.9%自報
Toolathlon50.0%自報
Finance Agent v244.9%

Code

Claw-Eval80.9%自報
FrontierSWE27.0%

Math

AIME 202696.4%自報
MathVision93.2%自報
HMMT Feb 2692.7%自報
IMO-AnswerBench86.0%自報
LiveBench72.2%

Multimodal

OSWorld-Verified73.1%自報

Reasoning

GPQANYU + Cohere + Anthropic (2023)90.5%自報
LiveCodeBench v689.6%自報
CharXiv-R86.7%自報
BrowseCompOpenAI (2025)86.3%自報
SWE-Bench Verified80.2%自報
SWE-bench Multilingual76.7%自報
Terminal-Bench 2.0Stanford × Laude Institute (2026)66.7%自報
OJBench60.6%自報
SWE-Bench ProPrinceton NLP (2024)58.6%自報
SciCode52.2%自報
Humanity's Last Exam36.4%自報
APEX-Agents27.9%自報

Search

DeepSearchQA83.0%自報
WideSearch80.8%自報

Vision

V*96.9%自報
MMMU-Pro80.1%自報
BabyVision68.5%自報

AA 評測指數

(Artificial Analysis)

暫無 AA 評測資料

LLM Stats 分類評分

(LLM Stats (zeroeval))
Math
80
Multimodal
80
Search
80
Frontend Development
80
Vision
80
Physics
70
Reasoning
70
General
70
Biology
70
Chemistry
70
Agents
60
Code
60
Tool Calling
60
Finance
40

定價

輸入價格$0.95 / 1M tokens
輸出價格$4 / 1M tokens
混合價格(3:1)$1.4375 / 1M tokens
快取讀取價格$0.16 / 1M tokens

速度

Tokens/秒38.6
首Token延遲2.85s
首回答延遲2.85s

供應商價格排行

供應商價格排行

25 個供應商

最便宜: JZSTOKEN最貴: Privatemode AI
供應商輸入輸出
1JZSTOKEN最便宜
$0.1557
$0.7784
2routing.run
$0.275
$1.1
3Cortecs
$0.475
$2.97
4Melious
$0.669
$4.014
5DeepInfra
$0.75
$3.5
6GreenPT
$0.7524
$4.275
7DevPass (LLM Gateway)
$0.8
$3.4
8Alibaba (China)
$0.929
$3.858
9Kimi主要
$0.95
$4
10Fireworks
$0.95
$4
11Novita
$0.95
$4
12Moonshot AI
$0.95
$4
13302.AI
$0.95
$4
14OpenCode
$0.95
$4
15Ollama Cloud
$0.95
$4
16AIHubMix
$0.95
$4
17Azure Cognitive Services
$0.95
$4
18Requesty
$0.95
$4
19DigitalOcean
$0.95
$4
20Azure
$0.95
$4
21Auriko
$0.95
$4
22Moonshot AI (China)
$0.9711
$4.0337
23DotHoosh
$1.21917
$5.13333
24UnoRouter
$1.2675
$5.3368
25Privatemode AI
$1.791
$8.9436

比較該模型在不同 API 供應商之間的定價。

外部連結