Qwen 3.8 27B (Colombo) vs Qwen 3.8 27B
The same model, run in two places. Qwen 3.8 27B (Colombo) runs on our own server in Sri Lanka, so prompts and answers never leave the country; the standard Qwen 3.8 27B runs abroad, with a longer context and fallback routes if one has a problem. Choose by where your data is allowed to go.
Choose Qwen 3.8 27B (Colombo) if
- Your data has to stay in Sri Lanka.
- Your users are in Sri Lanka and you want the model close to them.
- Low reasoning effort by default — on our server it scored higher that way.
Choose Qwen 3.8 27B if
- You need the full 262K-token context.
- You want automatic failover to another route if one has a problem.
- Your data has no requirement to stay in the country.
Side by side
| Qwen 3.8 27B (Colombo) | Qwen 3.8 27B | |
|---|---|---|
| Made by | Alibaba | Alibaba |
| Released | August 2026 | August 2026 |
| Input price, per 1M tokens | S$0.39 | S$0.5915 |
| Output price, per 1M tokens | S$1.95 | S$3.55 |
| Cached input, per 1M tokens | — | S$0.1521 |
| Batch (about half price) | No | No |
| Context window | 131K tokens | 262K tokens |
| Longest answer | — | 131K tokens |
| Reads | Text | Text, Images, Video |
| Thinks before answering | Yes, adjustable | Yes, adjustable |
| Open weights | Yes, Apache 2.0 | Yes, Apache 2.0 |
Prices are Roar AI’s published rates, live from our price list. Specs are each lab’s own published figures. Prices are our published US dollar rates converted at S$1.30 to the dollar, our rate for October — the same rate your card is charged.
What a real job costs
Take 1,000 requests, each about 3,000 tokens in and 1,000 out — a long prompt and a page of answer.
Qwen 3.8 27B (Colombo) does the same job for 1.7× less. Models that think before answering spend extra output tokens doing it, so on hard prompts the real gap can be wider than the rates suggest.
Use both, switch any time
On Roar AI both are on the same key and the same monthly invoice, paid in your local currency. Trying the other one is a one-word change:
client.chat.completions.create(model="qwen3.8-27b-lk", …) client.chat.completions.create(model="qwen3.8-27b", …)
Send the same prompts to both for a day and compare the answers on your own work — that settles it faster than any benchmark. Read the docs.
More head-to-heads
Try both
on one key
Send the same prompts to each and compare the answers on your own work.