Skip to main content
leaderboard
.cn
Toggle Sidebar
Search models or organizations…
⌘K
Explore
Frontier Models
Historical Trends
Local Models
Compare Models
Ranking Studio
Library
Models
Pricing
Organizations
Benchmarks
Data sources
Settings
Toggle Sidebar
leaderboard
.cn
Loading page content
Feedback
Loading page content
Back to models
Kimi-K2.5
Open
Share
Add to comparison
Tool use · Reasoning
Overview
Deployment
Pricing
Evaluations
Evaluation overview
Overall rank
#48
75 points · 180 models
Open-weight rank
#14
83.5 points · 109 models
Data coverage
82 items
9 sources · updated through 2026-07-16
Domain performance
Coding
#46 / 160
75.5
Research
#43 / 174
87.3
Agents
#52 / 125
70.1
Vision
#32 / 65
80.6
Long context
#31 / 82
78.2
Evaluation highlights
All evaluations
MMLU-Redux
Reasoning · Qwen3.6-Plus Technical Report
94.5%
9
/ 102
VideoMMMU
Multimodal · Qwen3.6-Plus Technical Report
86.6%
3
/ 24
LongBench v2
Long context · Qwen3.6-Plus Technical Report
61%
6
/ 22
SWE-bench Verified
Coding · Kimi K2.6 Technical Report
76.8%
22
/ 60
HLE (w/ Tools)
Agents · GLM-5.1 Technical Report
51.8%
14
/ 37
C-Eval
Chinese · Qwen3.6-Plus Technical Report
94%
1
/ 64
Basic information
Context (input / output)
262K / 16K
Parameters
1000B / 32B active
Release date
2026-01
License
Modified MIT
Input modalities
Text · Image · Video
Output modalities
Text
Active parameters
32B
Availability
Moonshot AI
Full pricing
Input
≈ $0.5714 /1M
Output
≈ $1.7143 /1M
vLLM · INT4
Deployment options
INT4 · 8 × 141 GB VRAM
Related links
Official site
HuggingFace
Version relationships
Kimi K2 family
Kimi K2 Base
2025-07 · Same series
Kimi-K2
2025-07 · Same series
Kimi-K2.5
Current
2026-01 · Current model
Kimi-K2.6
2026-04 · Same series
Kimi-K2.7-Code
2026-06 · Same series
Provider variants
Instant
Runtime modes
Thinking