Kimi K2.5
MoonshotAiFlagship
ThinkingTool UseVisionStructured Output
About this model
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 visual and text tokens, it delivers strong performance in general reasoning, visual coding, and agentic tool-calling.
Performance Tier
Flagship
Kimi K2.5 is a flagship model from MoonshotAi : the most capable in their lineup.
Best-in-class model from this provider. Highest performance across benchmarks, ideal for demanding tasks.
Pricing
This model is included in Elosia plans| Type | per 1M tokens |
|---|---|
| Input (prompt) | $0.420 |
| Output (completion) | $2.20 |
| Cache read | $0.210 |
Capabilities
Context Length262K
Max Output Tokens66K
TokenizerOther
Inputtext, image
Outputtext
Release DateJanuary 27, 2026
Benchmarks
General Intelligence
MMLU
85.8%
MMLU-Pro
87.1%
GPQA Diamond
62%
Mathematics
MATH-500
82.5%
AIME 2025
96.1%
Programming
HumanEval
86%
SWE-bench Verified
53%
Reasoning
IFEval
84.5%
Humanity's Last Exam
50.2%
Recommended Use Cases
General ChatCodingAnalysisCreative Writing
Strengths
- Strong general-purpose model from Moonshot AI
- Good math and coding performance for its price
- Long context window for document processing
- Strong Chinese and multilingual capabilities
Limitations
- Smaller ecosystem compared to Western AI providers
- Less extensive documentation in English
Resources
This model may use your data for training
