Kimi K2.6 Now on SiliconFlow: SOTA Long-horizon Coding

23 апр. 2026 г.

Содержание

TL; DR: Kimi K2.6 is now available on SiliconFlow. This open-source multimodal model delivers state-of-the-art long-horizon coding, autonomous agent orchestration, and coding-driven design capabilities. With 58.6 on SWE-Bench Pro and 86.3 on BrowseComp (Agent Swarm), K2.6 outperforms leading closed-source models in agentic workflows. Start building with SiliconFlow's API today to explore long-horizon coding tasks.

Overview: SOTA Coding Engine

Kimi K2.6, Moonshot AI's latest open-source multimodal agentic model.

This release brings exceptional long-horizon coding capabilities, proactive autonomous execution, and swarm-based task orchestration, enabling developers to achieve production-grade agentic workflows with greater efficiency and reliability.

As a next-generation Mixture-of-Experts (MoE) model with 1T total parameters and 32B activated parameters, Kimi K2.6 sets a new benchmark for open-source agentic AI.

Try Kimi K2.6 on SiliconFlow

Cost-effective Pricing: $0.16 / $0.95 / $4.00 per 1M tokens (Cached Input / Input / Output).
Seamless Integration: Instant compatibility with your existing development ecosystem: deploy via SiliconFlow's OpenAI-Compatible API through Cline, Gen-CLI, Kilo Code, Roo Code ; Anthropic-Compatible API with Claude Code; plug into agents like OpenClaw, Hermes Agent; ready-to-use in Dify, Janitor AI, Chub AI, ChatHub, Chatbox, Sider; and also available through OpenRouter.

Core Capabilities

Long-Horizon Coding Excellence: Sustains 4,000+ tool calls across 12+ hours of continuous execution, with robust generalization across Rust, Go, Python, and performance optimization tasks.
Agent Swarm Orchestration: Scales to 300 sub-agents executing 4,000 coordinated steps simultaneously, delivering end-to-end outputs from documents to websites in a single autonomous run.
Coding-Driven Design: Transforms simple prompts into production-ready interfaces with structured layouts, interactive elements, rich animations, and full-stack workflows.
Proactive Autonomous Execution: Powers persistent 24/7 background agents that manage schedules, execute code, and orchestrate cross-platform operations without human oversight.
Motion-rich frontend: Videos in hero sections, WebGL shaders, GSAP + Framer Motion, Three.js 3D.

Benchmark Performance

SWE-Bench Pro: 58.6 (vs. GPT-5.4: 57.7, Claude Opus 4.6: 53.4, Gemini 3.1 Pro: 54.2)
BrowseComp (Agent Swarm): 86.3 (vs. GPT-5.4: 82.7, Claude Opus 4.6: 83.7, Gemini 3.1 Pro:85.9)
DeepSearchQA (F1-score): 92.5 (vs. GPT-5.4: 78.6, Claude Opus 4.6: 91.3, Gemini 3.1 Pro: 81.9)
Terminal-Bench 2.0: 66.7 (vs. GPT-5.4 & Claude Opus 4.6: 65.4, Gemini 3.1 Pro: 68.5)

Kimi K2.6 consistently outperforms or matches leading closed-source models in agentic workflows, coding tasks, and long-context reasoning.

Real-World Applications

Autonomous Software Engineering: In one demonstration, K2.6 autonomously optimized an 8-year-old financial matching engine over 13 hours, executing 1,000+ tool calls to modify 4,000+ lines of code, achieving 185% throughput improvement.
Multi-Agent Research & Content Creation: Agent Swarm coordinates 300 specialized sub-agents to deliver comprehensive research reports, executive presentations, structured datasets, and custom visualizations in a single autonomous run.
Coding-Driven UI/UX Generation: Transforms simple prompts into production-ready front-end interfaces with deliberate aesthetic choices, interactive elements, scroll-triggered animations, and lightweight full-stack workflows.

Show Case

Promt: create a website of The best 100 books

Get Started Immediately

Explore: Try Kimi K2.6 in the SiliconFlow playground, and explore the Kimi Family Models, such as Kimi K2.5, Kimi-K2-Instruct-0905, Kimi-K2-Instruct.
API-reference: Explore the full API specifications in the SiliconFlow API documentation.

import requests

url = "https://api.siliconflow.com/v1/chat/completions"

payload = {
    "model": "moonshotai/Kimi-K2.6",
    "messages": [
        {
            "role": "user",
            "content": "Tell me a story"
        }
    ],
    "stream": True,
    "temperature": 1,
    "top_p": 0.95
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

import requests

url = "https://api.siliconflow.com/v1/chat/completions"

payload = {
    "model": "moonshotai/Kimi-K2.6",
    "messages": [
        {
            "role": "user",
            "content": "Tell me a story"
        }
    ],
    "stream": True,
    "temperature": 1,
    "top_p": 0.95
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

import requests

url = "https://api.siliconflow.com/v1/chat/completions"

payload = {
    "model": "moonshotai/Kimi-K2.6",
    "messages": [
        {
            "role": "user",
            "content": "Tell me a story"
        }
    ],
    "stream": True,
    "temperature": 1,
    "top_p": 0.95
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.text)