Kimi K2.6 (Non-reasoning)
KimiKimiOpen WeightModified 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
38
general
39
coding
79
reasoning
53
science估算
60
agents
80
multimodal
Science 在缺少專門科學評測時使用推理能力代理估算。
排行榜排名
基準測試分數 (LLM Stats)
Agents
BrowseComp
86.3%自報
DeepSearchQA
83.0%自報
Claw-Eval
80.9%自報
WideSearch
80.8%自報
OSWorld-Verified
73.1%自報
Terminal-Bench 2.0
66.7%自報
SWE-Bench Pro
58.6%自報
MCP-Mark
55.9%自報
Toolathlon
50.0%自報
APEX-Agents
27.9%自報
Biology
GPQA
90.5%自報
SciCode
52.2%自報
Code
SWE-Bench Verified
80.2%自報
SWE-bench Multilingual
76.7%自報
General
LiveCodeBench v6
89.6%自報
MMMU-Pro
80.1%自報
Math
AIME 2026
96.4%自報
MathVision
93.2%自報
HMMT Feb 26
92.7%自報
IMO-AnswerBench
86.0%自報
Humanity's Last Exam
36.4%自報
Multimodal
V*
96.9%自報
CharXiv-R
86.7%自報
BabyVision
68.5%自報
Reasoning
OJBench
60.6%自報
AA 評測指數
Intelligence Index43.0
Coding Index38.4
Tau20.9
Gpqa0.8
Lcr0.6
Ifbench0.4
Scicode0.4
Terminalbench Hard0.4
Hle0.2
LLM Stats 分類評分
Vision80
Coding80
Frontend Development80
General80
Math80
Multimodal80
Reasoning80
Search80
Agents70
Biology70
Chemistry70
Code70
Physics70
Tool Calling60
定價
輸入價格$0.95 / 1M tokens
輸出價格$4 / 1M tokens
混合價格(3:1)$1.712 / 1M tokens
速度
Tokens/秒28.5 tokens/s
首Token延遲1.71s
首回答延遲1.71s
可用提供商
(LS 內部計價單位)| 提供商 | 輸入價格 | 輸出價格 |
|---|---|---|
| Novita | 950K | 4.0M |
| Moonshot AI | 950K | 4.0M |