moonshotai/kimi-k2.6
Data publikacji: 2026-04-20
262,144 context · $0.95/M input tokens · $4.00/M output tokens
Kimi K2.6 is Moonshot AI’s open-source native multimodal agentic model, designed for long-horizon coding, coding-driven UI/UX generation, proactive autonomous execution, and multi-agent orchestration. Built on a 1T-parameter Mixture-of-Experts architecture with 32B active parameters, it supports text and image inputs, a 262K-token context window, thinking mode, preserve-thinking workflows, function calling, and structured outputs. It is especially strong for complex end-to-end coding tasks across Python, Rust, Go, front-end engineering, DevOps, performance optimization, and agentic workflow automation.
Cennik
Płać za użycie
Bez kosztów początkowych, płacisz tylko za to, czego używasz
Wypróbuj model
Użycie API
Użyj poniższych przykładów kodu, aby zintegrować się z naszym API:
import OpenAI from 'openai';
if (!process.env.WAVESPEED_API_KEY) throw new Error('Set WAVESPEED_API_KEY');
const client = new OpenAI({
apiKey: process.env.WAVESPEED_API_KEY,
baseURL: 'https://llm.wavespeed.ai/v1',
timeout: 120_000,
maxRetries: 2,
});
try {
const response = await client.chat.completions.create({
model: 'moonshotai/kimi-k2.6',
messages: [{ role: 'user', content: 'Hello!' }],
});
console.log(response.choices[0]?.message?.content ?? '');
} catch (error) {
console.error('LLM request failed:', error);
process.exitCode = 1;
}import OpenAI from 'openai';
if (!process.env.WAVESPEED_API_KEY) throw new Error('Set WAVESPEED_API_KEY');
const client = new OpenAI({
apiKey: process.env.WAVESPEED_API_KEY,
baseURL: 'https://llm.wavespeed.ai/v1',
timeout: 120_000,
maxRetries: 2,
});
try {
const response = await client.chat.completions.create({
model: 'moonshotai/kimi-k2.6',
messages: [{ role: 'user', content: 'Hello!' }],
});
console.log(response.choices[0]?.message?.content ?? '');
} catch (error) {
console.error('LLM request failed:', error);
process.exitCode = 1;
}Wprowadzenie do modelu
MoonshotAI: Kimi K2.6
Kimi K2.6 is Moonshot AI’s open-source native multimodal agentic model, designed for long-horizon coding, coding-driven UI/UX generation, proactive autonomous execution, and multi-agent orchestration. Built on a 1T-parameter Mixture-of-Experts architecture with 32B active parameters, it is optimized for complex coding, visual understanding, tool use, and large-scale agent workflows.
Why It Looks Great
- Open-source native multimodal agentic model from Moonshot AI
- 1T-parameter Mixture-of-Experts architecture with 32B active parameters
- 262K-token context window for long prompts, large codebases, documents, and multi-turn workflows
- Strong long-horizon coding performance across Python, Rust, Go, front-end, DevOps, and optimization tasks
- Excellent fit for coding-driven UI/UX generation, including full-stack apps and polished interfaces
- Agent Swarm capabilities for decomposing and coordinating complex multi-agent workflows
- Vision input support for screenshots, mockups, diagrams, and multimodal document understanding
- Thinking mode and preserve-thinking support for multi-step reasoning and coding agent scenarios
- Function calling and tool-use support for agentic application workflows
- Structured output support for JSON responses and schema-constrained generation
Key Features
- Architecture: Mixture-of-Experts
- Total Parameters: 1T
- Active Parameters: 32B
- Context Window: 262,144 tokens
- Max Input: Not listed
- Max Output: Not listed
- Input: Text, Image
- Output: Text
- Vision: Supported
- Function Calling: Supported
- Structured Outputs: Supported
- Thinking Mode: Supported
- Preserve Thinking: Supported
- Image Generation: Not listed
- Audio Input: Not listed
- Supported Parameters: frequency_penalty, include_reasoning, logit_bias, logprobs, max_tokens, min_p, parallel_tool_calls, presence_penalty, reasoning, reasoning_effort, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_p
Specifications
| Specification | Value |
|---|---|
| Provider | moonshot |
| Model Type | Chat Completions model |
| Architecture | Mixture-of-Experts |
| Parameters | 1T total / 32B active |
| Experts | 384 experts, 8 selected per token |
| Attention | MLA |
| Vision Encoder | MoonViT |
| Context Window | 262,144 tokens |
| Input | Text, Image |
| Output | Text |
| Vision | Supported |
| Function Calling | Supported |
| Structured Outputs | Supported |
| Thinking Mode | Supported |
Pricing
| Token Type | Cost |
|---|---|
| Input | $0.73 per million tokens |
| Output | $3.49 per million tokens |
| Cached Input | $0.25 per million tokens |
How to Use
- Write your prompt - describe the task, provide context, and specify the desired output format.
- Submit - the model processes your request and returns the response.
API Integration
Base URL: https://llm.wavespeed.ai/v1
API Endpoint: chat/completions
Model ID: moonshotai/kimi-k2.6
API Usage
Python SDK
from openai import OpenAI
client = OpenAI(
api_key="YOUR_API_KEY",
base_url="https://llm.wavespeed.ai/v1"
)
response = client.chat.completions.create(
model="moonshotai/kimi-k2.6",
messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)cURL
curl https://llm.wavespeed.ai/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "moonshotai/kimi-k2.6",
"messages": [{"role": "user", "content": "Hello!"}]
}'Notes
- Model: moonshotai/kimi-k2.6
- Provider: moonshot
- Best suited for long-horizon coding, UI/UX generation, visual understanding, tool use, multi-agent orchestration, and autonomous workflow execution
Info
Obsługiwane funkcje
Przewodnik dostępu do API
Kimi K2.6 API
moonshotai/kimi-k2.6
Kimi K2.6 is Moonshot AI’s open-source native multimodal agentic model, designed for long-horizon coding, coding-driven UI/UX generation, proactive autonomous execution, and multi-agent orchestration. Built on a 1T-parameter Mixture-of-Experts architecture with 32B active parameters, it supports text and image inputs, a 262K-token context window, thinking mode, preserve-thinking workflows, function calling, and structured outputs. It is especially strong for complex end-to-end coding tasks across Python, Rust, Go, front-end engineering, DevOps, performance optimization, and agentic workflow automation.
Wejście
$0.95 /M
Wyjście
$4 /M
Kontekst
262K
Maks. wyjście
262K
Vision
Obsługiwane
Użycie narzędzi
Obsługiwane
Wypróbuj Kimi K2.6 w WaveSpeedAI
Uzyskaj dostęp do Kimi K2.6 przez nasze ujednolicone API — kompatybilne z OpenAI, bez zimnych startów, przejrzyste ceny.
Najczęstsze pytania o Kimi K2.6
Ile kosztuje API Kimi K2.6?+
Cennik na WaveSpeedAI: $0.95 za milion tokenów wejściowych i $4.00 za milion tokenów wyjściowych. Prompt caching i przetwarzanie wsadowe są rozliczane oddzielnie i obniżają efektywny koszt długich, powtarzalnych obciążeń.
Jakie jest okno kontekstu Kimi K2.6?+
Kimi K2.6 obsługuje do 262K tokenów kontekstu i do 262K tokenów wyjściowych na zapytanie.
Czy Kimi K2.6 jest kompatybilny z OpenAI?+
WaveSpeedAI udostępnia Kimi K2.6 pod adresem https://llm.wavespeed.ai/v1 przez interfejs Chat Completions zgodny z OpenAI. W większości klientów OpenAI SDK wystarczy zmienić base URL i klucz API; opcjonalne pola zależą od modelu.
Jak zacząć z Kimi K2.6?+
Zaloguj się do WaveSpeedAI, utwórz klucz API w Access Keys i wyślij żądanie do https://llm.wavespeed.ai/v1/chat/completions z identyfikatorem modelu pokazanym powyżej. Bieżącą dostępność, możliwości i ceny sprawdź w katalogu modeli.