Issues / #770

#770 Qwen 3.8 27B on Strata?

closed · @frederikhors · 3 comments · View on GitHub

Models & quants

Description

_This project is amazing._

Given that the **quantized Qwen 3.8-Flash-Next** isn't all that much more powerful than **Qwen 3.8 27B** model, and considering that llama.cpp doesn't match Strata's optimization and speed, I wonder what kind of prompt processing and token generation speeds a Strata engine for Qwen 3.8 27B could achieve.

Have you thought about that?

Is there even a remote possibility that Strata could also run Qwen 3.8 27B?

Related on strata.com

Editorial links to help you install, pick models, or read release notes — not part of the upstream thread.