Kimi K3
Kimi K3 is a multimodal reasoning model developed by Moonshot AI. It supports text, image, and video inputs and is designed for complex tasks like coding, knowledge work, and long-horizon agentic workflows.
Kimi K3 specs and pricing
- Provider: MoonshotAI
- Released: July 16, 2026
- Context window: 1M tokens
- Max output: 944K tokens
- Input: text, images, video
- Output: text
- Price: $0.19 input / $20 output per 1M tokens
- Reasoning: yes
- Tool calling: yes
What Kimi K3 is good at
- Analyzing long documents and complex codebases
- Processing multimodal data including images and video
- Executing multi-step agentic workflows
To-dos to try with Kimi K3
- Analyze this 500-page technical manual and summarize the troubleshooting steps for the hardware.
- Review this video recording of a meeting and extract the key action items and assigned deadlines.
- Debug this multi-file software project and suggest a fix for the identified memory leak.
FAQ
- What types of media can Kimi K3 process? The model is multimodal and can accept text, images, and video as input.
- What is Kimi K3 primarily designed to do? It is built for complex coding, knowledge-intensive tasks, and long-horizon agentic workflows that require advanced reasoning.
- How much does Kimi K3 cost? Kimi K3 costs $0.19 per million input tokens and $20 per million output tokens through OpenRouter. In todo.is a typical agent step costs about 3.34 credits ($1 = 100 credits).
- What is the context window of Kimi K3? Kimi K3 can read up to 1M tokens at once and write up to 944K tokens in one reply.
- Can I use Kimi K3 without coding? Yes. In todo.is you pick Kimi K3 as your agent's model and write to-dos in plain words, on the web, the Android app, WhatsApp or Telegram. No API key needed: credits are included, or you can bring your own key.
JavaScript is required to use the todo.is app.