Qwen releases Qwen3.8-Flash-Next, a 125B a6B multimodal model

Qwen3.8-Flash-Next is a new multimodal model with 125B total parameters and 6B active (a6B), released on Hugging Face and ModelScope. It supports image-text-to-text tasks and is available in FP8 and GGUF quantizations, with day-0 support from unsloth.