qwen3-vl-moe & qwen3-omni-moe

[weiyangdaren]

Thank you for the amazing work!
I would like to know if there is any plan or timeline for supporting Qwen3-VL-MoE and Qwen3-Omni-MoE models. Could you please provide any updates on when these might be available?

[yghstill]

@weiyangdaren The quantization for Qwen3-VL-MoE and Qwen3-Omni-MoE will be released within two weeks, and support for speculative decoding is expected to be available in about a month.