Qwen3.8-27B (Colombo)
Qwen3.8-27B on our own server in Colombo: your data is processed in Sri Lanka and nowhere else.
About Qwen3.8-27B (Colombo)
This is Qwen3.8-27B — Alibaba’s compact open model, strong at coding and documents for its size — running on our own GPU server in Colombo. Prompts and answers are processed in Sri Lanka and never leave it: there is no fallback to a server abroad, by design.
We run it with a 131K-token context and low reasoning effort by default. On 250 MMLU-Pro questions through this server, that setting scored 84.0% — higher than the model’s own default, using a third of the tokens. A single request streams at about 73 tokens a second, and it supports tool calling.
Where it’s strong
Work whose data has to stay in Sri Lanka.
Drafting, summaries and question answering for local teams.
Agents that call your own tools, kept inside the country.
Where to be careful
One server and no fallback abroad: during maintenance or an outage, requests fail rather than leave the country.
A 131K-token context here, shorter than the 262K the model supports elsewhere.
Price
For scale: 1,000 requests of about 3,000 tokens in and 1,000 out cost $2.40 — a long prompt and a page of answer each. Every rate is published and stays the same from one request to the next. See every model.
Specs
qwen3.8-27b-lkCall it
Use the OpenAI SDK you already have — point it at Roar AI and name the model. Read the docs.
from openai import OpenAI
client = OpenAI(base_url="https://api.roar-ai.com/v1", api_key="roar_live_…")
reply = client.chat.completions.create(
model="qwen3.8-27b-lk",
messages=[{"role": "user", "content": "Summarise this contract in plain English."}],
)Compare Qwen3.8-27B (Colombo)
Questions
How much does Qwen3.8-27B (Colombo) cost?
Input is $0.30 per 1M tokens; output is $1.50 per 1M tokens. For scale, 1,000 requests of about 3,000 tokens in and 1,000 out cost $2.40. The rate is published and does not change from one request to the next.
Can I use Qwen3.8-27B (Colombo) alongside other models?
Yes. Every model on Roar AI is on the same API key and the same monthly invoice, so switching from Qwen3.8-27B (Colombo) to another model is a change to one word in your code.
Is Qwen3.8-27B (Colombo) open source?
Its weights are open under the Apache 2.0 licence, so you could run it yourself. Through Roar AI you call it like any other model, with no servers to manage.