Kimi K2.5 (Non-reasoning)
KimiKimiOpen WeightMIT · Commercial OK
Description
Kimi K2.5 is Moonshot AI's flagship agentic model and a new SOTA open model. It unifies vision and text, thinking and non-thinking modes, and single-agent and multi-agent execution into one model. Built with Full-Parameter RL tuning, it achieves state-of-the-art performance across agents, coding, image, and video benchmarks.
Release Date
2026-01-27
Parameters
1.0T
Context Length
262K
Modalities
image, text, video
Capability Radar
18
general
50
coding
79
reasoning
59
science
50
agents
80
multimodal
Rankings
| Domain | #Rank | Score | Source |
|---|---|---|---|
| Code Ranking | 287 | 53.0 | AA |
| General Ranking | 257 | 47.0 | AA |
| Multimodal Ranking | 23 | 62.0 | LS |
| Science | 263 | 50.0 | AA |
Benchmark Scores (LLM Stats)
(LLM Stats (zeroeval))Code
PaperBench
63.5%SR
CyberGym
41.3%SR
FrontierSWE
26.0%
Finance
FinSearchComp T2&T3
67.8%SR
Language
MMLU-Pro
87.1%SR
Long Context
LongBench v2
61.0%SR
Math
AIME 2025
96.1%SR
HMMT 2025
95.4%SR
MathVista-Mini
90.1%SR
MathVision
84.2%SR
IMO-AnswerBench
81.8%SR
LiveBench
69.1%
Multimodal
Video-MME
87.4%SR
VideoMMMU
86.6%SR
Reasoning
GPQANYU + Cohere + Anthropic (2023)
87.6%SR
LiveCodeBench v6
85.0%SR
CharXiv-R
77.5%SR
SWE-Bench Verified
76.8%SR
BrowseCompOpenAI (2025)
74.9%SR
SWE-bench Multilingual
73.0%SR
AA-LCR
70.0%SR
OJBench (C++)
57.4%SR
Seal-0
57.4%SR
Terminal-Bench 2.0Stanford × Laude Institute (2026)
50.8%SR
SWE-Bench ProPrinceton NLP (2024)
50.7%SR
Humanity's Last Exam
50.2%SR
SciCode
48.7%SR
Search
WideSearch
79.0%SR
DeepSearchQA
77.1%SR
Vision
InfoVQAtest
92.6%SR
OCRBench
92.3%SR
OmniDocBench 1.5
88.8%SR
MMVU
80.4%SR
LongVideoBench
79.8%SR
MMMU-Pro
78.5%SR
LVBench
75.9%SR
SimpleVQA
0.71 / 100SR
MotionBench
70.4%SR
WorldVQA
46.3%SR
ZEROBench
0.11 / 100SR
AA Evaluation Indices
(Artificial Analysis)Tau2(Sierra + U Toronto + Vector Institute (2025))81.3
Gpqa(NYU + Cohere + Anthropic (2023))78.9
Lcr(Artificial Analysis)67.3
Ifbench(Google Research (2023))43.7
Intelligence Index(Artificial Analysis)19.4
Terminalbench Hard(Stanford × Laude Institute (2026))18.9
Hle(Center for AI Safety + Scale AI (2025))13.2
LLM Stats Category Scores
(LLM Stats (zeroeval))Language90
Legal90
Finance90
Image To Text80
Long Context80
Math80
Frontend Development80
Video80
Multimodal70
Physics70
Reasoning70
Search70
Structured Output70
General70
Healthcare70
Biology70
Chemistry70
Vision70
Agents60
Code50
Tool Calling50
Safety40
Pricing
Input Price$0.6 / 1M tokens
Output Price$3 / 1M tokens
Blended Price (3:1)$1.2 / 1M tokens
Cache Read Price$0.07 / 1M tokens
Speed
Tokens/sec0.0
Time to First Token0.00s
Time to Answer0.00s
Provider Price Ranking
Provider Price Ranking
11 providers
Cheapest: NanoGPTMost Expensive: Ofox
ProviderInputOutput
1NanoGPTCheapest
$0.3
$1.9
2OpenRouter
$0.45
$2.25
3TensorX
$0.5
$2.8
4ZenMux
$0.58
$3.02
5KimiPRIMARY
$0.6
$3
6Jiekou.AI
$0.6
$3
7NovitaAI
$0.6
$3
8Kilo Gateway
$0.6
$3
9Vercel AI Gateway
$0.6
$3
10HPC-AI
$0.6
$3
11Ofox
$0.6
$3
Compare pricing across different API providers for this model.