Qwen 3.8 - 27B
It was finally released so I tried it right away. Is it the (9t/s) setting that is unusually slow, or is the default size (8bit quantization model) mismatched to 128GB/m5max? In any case, I quickly shared the data as it was interesting. 1st LMstudio Simple Reasoning 2nd ollama+claudecode collaboration inference large…
This is Qwen 3.8 27B in the pi harness with oh-my-pi extension.
The trees and extra elements are much richer than vanilla pi but the pagoda is floating. The birds are a nice touch. oh-my-pi used 145k tokens (6.5x more than vanilla pi!) over 85 turns!
GLM 5.3 Max vs Fable 5 vs Qwen 3.8 Max vs Grok 4.6
Qwen 3.8 27B, Procedural Medieval City, I kind of like it.
You have reached the end of the archive
All of qwen38