Fastest & Best quality & Largest context tested
qwen3.8-27b-Swift1.5-Uncensored-ATX-AI3
27B
88.7 tok/sQuality 87.8Context 196,608 tokens
Highest token generation speed at 88.7 tok/s. Highest completed quality score at 87.8. Largest context window tested on this hardware, at 196,608 tokens.