[vllm-omni]: Omni Quant Support

[thuang6]

Feature Description

This epic is created to track Omni model (vllm-omni targeted) quantization support status.
main dtypes includes: INT4, MXFP8, MXFP4 is future target

Motivation and Use Case

omni model enable status:

model	architecture	owner	INT4	MXFP8
Qwen3-omni	Qwen3OmniMoeForConditionalGeneration	Liang	ok
Qwen2.5-omni	Qwen2_5OmniForConditionalGeneration	Liang	ok
GLM-Image	GlmImageForConditionalGeneration GlmImageTransformer2DModel	Liang	ok
BAGEL-7B-MoT	BagelForConditionalGeneration	Liang	ok
Ovis-Image	OvisImageTransformer2DModel	Liang	ok
HunyuanImage-3.0	HunyuanImage3ForCausalMM	Chang	ok
OmniGen2	-	Chang	WIP
Z-Image	ZImageTransformer2DModel	Xin	ok
Next-Step	NextStep	Xin	WIP (Feature request in vllm-omni)
LongCat-Image	LongCatImagePipeline	Xin	WIP
MiMo-Audio-7B-Instruct	MiMoAudioForCausalLM	Weiwei	ok
Qwen3-TTS-12Hz-1.7B-CustomVoice	Qwen3TTSForConditionalGeneration	Weiwei	ok
stable-audio-open	-	Weiwei	ok
FLUX.1-dev	FluxTransformer2DModel	Mengni	ok
SD	-	Mengni	-
Wan	WanTransformer3DModel	Mengni/Liang	ok
QwenImage	-	Mengni	-

vllm-omni Support Status

vllm-project/vllm-omni#1325

HF model id	Type	status	owner	PR
black-forest-labs/FLUX.1-dev	text→image	Merged	Yi	#1777
Qwen/Qwen3-Omni-30B-A3B-Instruct	text+audio→text+audio	Merged	Liang	#2670
Qwen/Qwen2.5-Omni-7B	text+audio→text+audio	Merged	Liang	#2670
Qwen/Qwen-Image	text→image	Merged	Yi	#3588
Wan-AI/Wan2.2-T2V-A14B-Diffusers	text→video	Merged	Liang	#3353
Wan-AI/Wan2.2-I2V-A14B-Diffusers	image→video	Merged	Liang	#3353
Wan-AI/Wan2.2-S2V-14B	audio+image+text→video	WIP	Liang
zai-org/GLM-Image	text→image	Merged	Liang	#3059

[mengniwang95]

#1519 diffusion model saving is done

[changwangss]

#1502 merged, HunyuanImage-3.0 is ready for auto-round.