POSTopenai
/v1/chat/completionsKimi K3
kimi-k3Kimi K3 is a high-performance AI model designed for advanced reasoning, coding, and complex task execution. It delivers strong performance on large-scale programming projects, multi-step problem solving, and knowledge-intensive applications. With broad context handling and multimodal capabilities, K3 is suitable for building powerful AI agents, developer tools, and enterprise-level applications.
Total Context
1Mtokens
Max Output
131.1Ktokens
Released
Jul 16, 2026
Modalities
Kimi K3 Price
| Input Price | Output Price | Cache Read |
|---|---|---|
| $2.8571/M | $14.2857/M | $0.2857/M |
Kimi K3 API
POSTanthropic
/v1/messagesKimi K3 Benchmark
Kimi K3
57.1
/100
Artificial Analysis Intelligence Index
Artificial Analysis broad capability aggregate
Index score
76.2
/100
Artificial Analysis Coding Index
Artificial Analysis software task aggregate
Index score
Knowledge & Reasoning
GPQA
Advanced science problem solving
93.5%
HLE
Broad expert-level exam set
44.3%
Coding & Engineering
SciCode
Scientific coding challenges
58.7%
Instruction Following & Agent Tasks
AA-LCR
Long-context reasoning
74.7%
Metrics sourced from Artificial Analysis
Kimi K3 FAQs
Common questions about using Kimi K3 on TokenHub.
What is Kimi K3?+
Kimi K3 is Moonshot AI's flagship multimodal reasoning model for complex coding, knowledge work and long-horizon agent tasks.
What are Kimi K3's notable capabilities?+
It combines native visual understanding, a context window of up to 1M tokens, advanced reasoning controls and strong long-horizon tool use.
What are the main tradeoffs of Kimi K3?+
Its large context and deep reasoning can increase latency and usage cost, and reliable agent runs require careful prompt, tool and context management.
How should I call Kimi K3 through TokenHub?+
Select Kimi K3 in TokenHub and follow the provider's current API settings; direct integrations commonly use the model ID k3 and an OpenAI-compatible request format.
Who should choose Kimi K3 over Kimi K2.6 or Kimi K2.7 Code?+
Choose it for demanding multimodal, long-context and agentic work; choose Kimi K2.6 for a more balanced general model or Kimi K2.7 Code for a focused coding workflow.
Kimi K3 Media, Demos and Reviews
Launch announcements, independent evaluations and community discussion about Kimi K3.
X (Twitter)
Reddit
YouTube