All models
Moonshot
Kimi K3
The new generation of Kimi with a huge context
Text
Images
Video
Reasoning
Tools
- Type
- Chat
- Context
- 1.0M tokens
- Russian
- medium
- Input price
- 20,32 ₽ / 1M tokens
- Output price
- 1 014,56 ₽ / 1M tokens
- Release date
- July 2026
chat/completions
import OpenAI from "openai"const client = new OpenAI({ baseURL: "https://api.yenisei.ru/v1", apiKey: process.env.YENISEI_API_KEY,})const stream = await client.chat.completions.create({ model: "kimi-k3", messages: [{ role: "user", content: "Hi!" }], stream: true,})for await (const chunk of stream) { process.stdout.write(chunk.choices[0]?.delta?.content ?? "")}About the model
Kimi K3 is the new generation of the Kimi lineup from Moonshot. It works with text and images: it analyzes pictures, screenshots and documents with visual elements. The model lays out detailed step-by-step reasoning for the most complex multi-stage tasks and is especially strong in planning and coordinating agent actions. Its very large context window fits huge documents and long conversations. It reliably calls tools (function-calling) and suits autonomous agents and automation. It is multilingual and its Russian is average.
Strengths
- The deep reasoning of the new Kimi generation
- Very large context
- Works with text and images
- Strong agent scenarios and planning
- Reliable tool calling (function-calling)
Where to use it
- Autonomous agents and multi-step planning
- Working with huge documents and correspondence
- Complex analytics and research
- Assistants that analyze images
- Agent scenarios with tool calling
- Working through tangled multi-stage tasks