Moonshot AI

MoonshotAI: Kimi K2.5

by Moonshot AItext+image->text5 endpoints
Neura Intelligence Index
71.9/ 100
Rank
#74
Confidence
high
Benchmarks
9(23% coverage)

Overview

Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability and a self-directed agent swarm paradigm. Built on Kimi K2 with continued pretraining over approximately 15T mixed...

Capabilities

Text generation
Image understanding
Tool use / Function calling
Structured output
Extended reasoning
Prompt caching
Long context
Top-P sampling
Stop sequences
Deterministic seed
Log probabilities

Modalities

Input
TextImage
Output
Text

Technical Specifications

Context Window
262.1K tokens
Max Output
235.9K tokens
Tokenizer
Other
Uptime (24h)
9994.1%
Parameters
1000000000000B
License
mitOpen Weights
Country
🇨🇳CN
Release Date
January 27, 20267mo

Supported Parameters

frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p

Benchmark Performance

coding

BenchmarkScoreSource
SWE-bench Verified
76.8
verified
SWE-Bench Pro
50.7
verified
SciCode
48.7
verified

math

BenchmarkScoreSource
AIME 2025
96.1
verified

reasoning

BenchmarkScoreSource
GPQA Diamond
87.6
verified
Humanity's Last Exam
50.2
verified

multimodal

BenchmarkScoreSource
MMMU-Pro
78.5
verified
CharXiv Reasoning
77.5
verified

agent

BenchmarkScoreSource
BrowseComp
74.9
verified

Performance Over Time

Pricing

Per 1M tokens
Input$0.45
Output$2.25
Cache read$0.07
Added
January 27, 2026
Last synced 9/20/2026