Kimi K3 API
kimi-k3Moonshot AICommercial use API verifiedText & reasoning
Frontier coding, native vision, and 1M-token agents from Moonshot AI.
by Moonshot AI
Pricing: Input 239 credits / 1M tokens (≈ $2.39), Output 1,195 credits / 1M tokens (≈ $11.95).
Save up to 28%with the $1,250 top-up pack (+10% bonus) — 20% lower than direct rates even without itTop up24H Status Monitor
Operational24h agonow
MODEL
kimi-k3
Tools
Reasoning effort
Thinking process
Max tokens
Start a conversation with Kimi K3
Input 239 · Output 1195 cr / 1M · billed per token
You only pay for successful generations — if a task fails or returns no result, its credits are refunded automatically, never charged. And your credits never expire: top up once and spend at your own pace.1 credit = $0.01 USDTop up $1,250 for +10% bonus credits → up to ~9% more off every call.Top up credits
Kimi K3 API pricing2 prices
Model & ModalityCredits / GenOur Price (USD)Direct / Standard Price (USD)Base savingsSavings with top-up$1,250 top-up · +10% credits
Kimi K3, Input
textMoonshot AI
Credits / Gen239per million tokens
Our Price (USD)$2.39
Direct / Standard Price (USD)$3
Base savings−20.3%
Savings with top-up · $1,250 · +10%−28% ↓
Kimi K3, Output
textMoonshot AI
Credits / Gen1,195per million tokens
Our Price (USD)$11.95
Direct / Standard Price (USD)$15
Base savings−20.3%
Savings with top-up · $1,250 · +10%−28% ↓
Explore more LLM models
Claude Fable 5.1
Anthropic's Claude Fable 5.1 — for demanding reasoning and long-horizon agentic work, with stronger long-running agentic coding, multistep research, and document, spreadsheet and slide work across a 1M-token context.
Claude Opus 5.5
Anthropic's Claude Opus 5.5 — frontier reasoning for complex coding, research and agentic work, served through one API.
Claude Sonnet 5.5
Anthropic's Claude Sonnet 5.5 — the balanced Claude for everyday coding, writing and analysis, served through one API.
Claude Haiku 5.5
Anthropic's Claude Haiku 5.5 — the fastest, lowest-cost Claude for high-volume chat, coding and everyday tasks, served through one API.
Gemini 3.6 Flash
Google's newest speed-optimized multimodal model — fast text, vision, and reasoning.
Gemini 3.7 Flash
Google's newest Flash model — a 1M-token context window and up to 65,536 tokens of output, at the same price as 3.6.