The Hugging Face repository lists variants spanning IQ2 and IQ3 quantization families, including versions with and without MTP. The smallest model file is 8.42 GB, reducing storage and memory requirements for local inference on consumer GPUs and unified-memory Macs. The release makes several packaging options available for local evaluation; the supplied repository evidence does not establish comparative speed, reasoning quality, or hardware compatibility beyond the stated intended use.
Raw SignalOct 12, 20261 min read1 source
UkisAI has released four mixed-precision GGUF versions of Swift 1.5, a post-trained Qwen3.8-27B reasoning model.
In one minute
UkisAI has published four mixed-precision GGUF variants of Swift 1.5, a post-trained Qwen3.8-27B reasoning model.