loopai-hq/Qwen3.8-27B-Pulsar
Text Generation • Updated
Fast Qwen3.8-27B and Qwen3.6-35B-A3B on Apple Silicon. Engine: github.com/loopai-hq/pulsar
Note Run with Pulsar 1.1.5: 153 tok/s writing code, 344 editing (M5 Max)
Note MoE model, run with Pulsar 1.1.5