Private, High-Performance
AI Inference

Private Qwen3.8 27B inference for developers and teams that need long-context, multimodal AI on Lithune-owned infrastructure in Türkiye. OpenRouter availability is pending.

OpenRouter Launch Pending → OpenRouter listing is pending; pricing and public access will be available there after launch.
Why Lithune

Regional infrastructure with clear data boundaries

Built for developers and teams that need capable multimodal inference, long context and a clearly documented data path.

Türkiye-hosted inference

Requests run on Lithune-owned hardware in Türkiye, without forwarding content to an external model or cloud GPU provider.

Long-context multimodal

Qwen3.8 27B combines text and image input, flexible reasoning and a 262,144-token context window behind an OpenAI-compatible API.

Provider-side privacy

Lithune does not retain prompts or completions or use them for training; provider boundaries and operational metadata are documented publicly.

Qwen3.8 27B AWQ INT4 · Text and image input
Context length262,144 tokens
Maximum output131,072 tokens
InfrastructureTürkiye · Private hardware
Vision & Multimodal Native image input (URL & Base64). Limits →
Prefix Caching Fast KV caching with prompt cache pricing discount
Provider-Side Zero Retention Prompt and completion content is not stored by Lithune or used for training. Data Policy →
Operations & Support

A transparent operating model

Inference runs on privately owned Lithune hardware in Türkiye. Product, privacy, legal and security questions share one monitored support channel.

Contact Lithune → Read data policy →
API Quickstart

OpenAI-compatible

Integration preview — keep these examples ready for use after the OpenRouter listing is published.

● ● ● 
from openai import OpenAI

client = OpenAI(
    base_url="https://openrouter.ai/api/v1",
    api_key="sk-or-...",
)

response = client.chat.completions.create(
    model="qwen/qwen3.8-27b",
    messages=[{"role": "user", "content": "Hello!"}],
    stream=True,
    extra_body={"provider": {"order": ["lithune"], "allow_fallbacks": False}},
)

for chunk in response:
    print(chunk.choices[0].delta.content or "", end="")
curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer sk-or-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3.8-27b",
    "messages": [{"role": "user", "content": "Hello!"}],
    "provider": {"order": ["lithune"], "allow_fallbacks": false},
    "stream": true
  }'
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://openrouter.ai/api/v1",
  apiKey: "sk-or-...",
});

const stream = await client.chat.completions.create({
  model: "qwen/qwen3.8-27b",
  messages: [{ role: "user", content: "Hello!" }],
  stream: true,
  provider: { order: ["lithune"], allow_fallbacks: false },
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content || "");
}

Lithune provides provider-side zero retention. OpenRouter and Cloudflare process data under their own policies; derived KV-cache blocks may remain temporarily in volatile memory until eviction or service restart.