LaunchAI ModelsJuly 4, 2026
Qwen3.5 INT8 text encoders released for ComfyUI
Community release of Qwen3.5 2B/4B/9B text encoders quantized to INT8 using ConvRot. Drop-in replacements for BF16, runs on 8GB VRAM.
Community release of Qwen3.5 2B/4B/9B text encoders quantized to INT8 using ConvRot. Drop-in replacements for BF16, runs on 8GB VRAM.