Native Multimodal Models for Vision Language Understanding and Generation

Chunjiang Ge Mingyang Han

Introduction

The unification of different deep learning architectures and tasks has reshaped various fields. The standardization of natural language processing via Large Language Models have revenlotionized NLP. This progress raises a critical question: Can we seamlessly integrate vision and language understanding with generation? Emerging native multimodal models like Gemini and GPT-4o demonstrate early successes in bridging these capabilities. While architectures for LLMs and vision-language models (e.g., LLaVA, Qwen-VL) show signs of convergence, vision generation remains fragmented across three paradigms: discrete autoregressive models, continuous diffusion models, and flow matching. This divergence highlights fundamental challenges in unifying multimodal understanding and generation. We systematically analyze existing approaches to identify optimal pathways for building truly unified multimodal architectures. We categoried these unified models into three approaches via the visual generation method: visual generation through external generator, discrete and continuous modeling.

Taxonomy of Unified Models

Benchmark Results

Understanding

Model	Params	POPE	MME-P	MMB_dev	SEED	VQAv2	GQA	MMMU	MM-Vet	TextVQA	MMStar
LWM		75.2	-	-	-	55.8	44.8	26.3	9.6	18.8	-
Unified-IO2	6.8B	-	-	71.5	-	-	-	86.2	-	-	61.8
LaVIT	7B	-	-	-	-	66.0	46.8	-	-	-	-
Emu	13B	-	-	-	-	52.0	-	-	-	-	-
Emu2	37B	-	1345	63.6	-	84.9	65.1	34.1	48.5	66.6	-
Emu3	8B	85.2	1243.8	58.5	68.2	75.1	60.3	31.6	-	64.7	-
SEEDLLaMA-I	8B	-	-	45.8	51.5	66.2	-	-	-	-	31.7
SEED-X	17B	84.1	1457.0	70.1	66.5	71.2	49.1	35.6	43.0	XXXXX	-
Janus	1.3B	87.0	1338.0	69.4	63.7	77.3	59.1	30.5	34.3	-	-
JanusFlow	1.3B	88.0	1333.1	74.9	70.5	79.8	60.3	29.3	30.9	-	-
Janus-Pro	1.5B	86.2	1444.0	75.5	68.3	-	59.3	36.3	39.8	-	-
Janus-Pro	7B	87.4	1567.1	79.2	72.1	-	62.0	41.0	50.0	-	-
NExT-GPT	13B	-	-	-	-	66.7	-	-	-	-	-
MUSE-VL	7B	-	1480.9	72.1	70.0	-	-	42.3	-	-	48.3
MUSE-VL	32B	-	1581.6	81.8	71.0	-	-	50.1	-	-	56.7
Libra	11.3B	88.2	1494.7	65.2	62.7	77.3	63.8	-	31.8	-	-
TokenFLow-XL	14B	87.8	-	76.8	72.6	77.6	62.5	43.2	-	62.3	-
QLIP	7B	86.1	1498.3	-	-	78.3	61.8	-	33.3	55.2	-
UniTok	7B	83.2	1448	-	-	76.8	61.1	-	33.9	51.6	-
DualToken	3B	86.0	1489.2	70.9	70.2	77.8	-	38.6	32.5	-	-
Liquid	7B	81.1	1119.3	-	-	71.3	58.4	-	-	42.4	-
SynerGen-VL	2.4B	85.3	1837	53.7	62.0	-	59.7	34.2	34.5	67.5	-
AnyGPT	7B	-	-	-	-	-	-	-	-	XXXXX	-
MIO-Instruct	7B	-	-	-	54.4	65.5	-	-	-	XXXXX	-
ILLUME	7B	88.5	1445.3	75.1	72.9	66.2	-	38.2	37.0	72.1	31.7
VL-GPT	7B	-	-	-	-	67.2	51.5	-	-	XXXXX	-
MM-Interleaved	13B	-	-	-	-	80.2	60.5	-	-	61.0	-
Gemini-Nano-1	1.8B	-	-	-	-	62.7	-	26.3	-	-	-
EasyGen	7B	-	-	-	-	-	44.6	-	-	XXXXX	-
DreamLLM	7B	-	-	58.2	-	72.9	-	-	36.6	-	-
DEEM-VQA	7B	-	-	60.8	-	68.2	55.7	-	37.4	XXXXX	-
X-VILA	7B	-	-	-	-	72.9	-	33.9	-	XXXXX	-
MetaMorph	8B	-	-	75.2	71.8	-	-	41.8	-	60.5	44.0
VILA-U	7B	85.8	1401.8	-	59.0	79.4	60.8	-	33.5	-	-
Chameleon	7B	-	170.0	31.1	30.6	-	-	25.4	8.3	-	31.1
Chameleon	30B	-	575.3	32.5	48.5	-	-	38.8	-	-	31.8
Video-LaVIT	7B	-	1551.8	67.3	64.0	-	-	-	-	-	-
Show-o	1.3B	84.5	1232.9	-	-	74.7	61.0	27.4	-	-	-
HermesFlow	1.3B	81.4	1249.7	-	-	75.3	61.7	28.3	-	-	-
Orthus	7B	79.6	1265.8	-	-	63.2	52.8	28.2	-	-	-
Liquid†	7B	81.1	1119.3	-	-	71.3*	58.4*	-	-	42.4	-
D-DiT	2B	84.0	1124.7	-	-	60.1	59.2	-	-	-	-
MMAR	7B	83.0	1393.9	66.32	64.5	-	-	-	27.8	-	-
LLaMAFusion	8B	-	1603.7	-	-	-	-	41.7	-	-	-
OmniMamba	1.3B	86.3	1290.6	-	-	77.7	60.8	30.6	-	-	-
ILLUME+	3B	87.6	1414.0	80.8	73.3	-	-	44.3	40.3	69.9	-
MetaQuery-XL	7B	-	1685.2	83.5	76.9	-	-	58.6	66.6	-	-
VARGPT	7B	87.3	1488.8	67.6	67.9	78.4	62.3	36.4	-	54.1	-
VARGPT-v1.1	7B	89.1	1684.1	81.01	76.0	80.4	66.2	48.5	-	82.0	-
UniToken	7B	-	-	71.1	69.9	-	-	32.8	-	-	46.1

MJHQ-30K

Method	Resolution	Params	#Images	FID
Generation Only
SD-XL		-	2000M	9.55
PixArt		-	25M	6.14
Playground v2.5		-	-	4.48
Unified Model
LWM		7B	-	17.77
Show-o		1.3B	36M	15.18
JanusFlow		1.3B	-	9.51
MUSE-VL	256	7B	30K	7.73
Janus		1.3B	-	10.10
VILA-U	256	7B	15M	12.81
VILA-U	384	7B	15M	7.69
SynerGen-VL		2.4B	30K	6.10
Liquid	512	7B	30M	5.47
ILLUME		7B	30K	7.76
ILLUME+		3B	30K	6.00
MetaQuery-XL		7B	30K	6.02

GenEval Bench

Model	Params	Res.	Single Obj.	Two Obj.	Count.	Colors	Position	Color Attri.	Overall↑
Generation Model
LlamaGen	0.8B	-	0.71	0.34	0.21	0.58	0.07	0.04	0.32
LDM	1.4B	-	0.92	0.29	0.23	0.70	0.02	0.05	0.37
PixArt-α	0.6B	-	0.98	0.50	0.44	0.80	0.08	0.07	0.48
VAR	-	256	-	-	-	-	-	-	0.53
Emu3-Gen	8B	-	0.98	0.71	0.34	0.81	0.17	0.21	0.54
SDv1.5	0.9B	-	0.97	0.38	0.35	0.76	0.04	0.06	0.43
SDv2.1	0.9B	-	0.98	0.51	0.44	0.85	0.07	0.17	0.50
SDXL	2.6B	-	0.98	0.74	0.39	0.85	0.15	0.23	0.55
SD3	2B	-	0.98	0.74	0.63	0.67	0.34	0.36	0.62
IF-XL	4.3B	-	0.97	0.74	0.66	0.81	0.13	0.35	0.61
DALL-E 2	6.5B	-	0.94	0.66	0.49	0.77	0.10	0.19	0.52
DALL-E 3	-	-	0.96	0.87	0.47	0.83	0.43	0.45	0.67
Unified Model
CODI	-	-	0.89	0.16	0.16	0.65	0.02	0.01	0.31
BSQViT	-	-	-	-	-	-	-	-	0.31
Chameleon	34B	-	-	-	-	-	-	-	0.39
LWM	7B	-	0.93	0.41	0.46	0.79	0.09	0.15	0.47
QLIP	7B	-	-	-	-	-	-	-	0.48
SEED-X	17B	-	0.97	0.58	0.26	0.80	0.19	0.14	0.49
MUSE-VL	7B	256	-	-	-	-	-	-	0.53
TokenFLow	13B	256	-	-	-	-	-	-	0.55
Orthus	7B	512	0.99	0.75	0.26	0.84	0.28	0.38	0.58
SynerGen-VL	2.4B	-	0.99	0.71	0.34	0.87	0.37	0.37	0.61
ILLUME	7B	-	0.99	0.86	0.45	0.71	0.39	0.28	0.61
ILLUME+	3B	-	0.99	0.88	0.62	0.84	0.42	0.53	0.72
Emu3-Gen	8B	-	0.98	0.71	0.34	0.87	0.37	0.37	0.61
Transfusion	-	256	-	-	-	-	-	-	0.63
D-DiT	2B	-	0.97	0.80	0.54	0.76	0.32	0.50	0.65
Show-o	1.3B	-	0.98	0.80	0.66	0.84	0.31	0.50	0.68
HermesFlow	1.3B	-	0.98	0.84	0.66	0.82	0.32	0.52	0.69
Janus	1.3B	-	0.97	0.68	0.30	0.84	0.46	0.42	0.61
JanusFlow	1.3B	-	0.97	0.59	0.45	0.83	0.53	0.42	0.63
Janus-Pro	1.5B	-	0.98	0.82	0.51	0.89	0.65	0.56	0.73
Janus-Pro	7B	-	0.99	0.89	0.59	0.90	0.79	0.66	0.80
MetaQuery-XL	7B	-	-	-	-	-	-	-	0.80
VARGPT-v1.1	7B	-	0.96	0.53	0.48	0.83	0.13	0.21	0.53
UniToken	7B	-	0.99	0.80	0.35	0.84	0.38	0.39	0.63

Image Tokenizer

Discrete Encoder/VQ

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
17/11	VQ-VAE	Neural Discrete Representation Learning	arXiv
19/06	VQ-VAE-2	Generating Diverse High-Fidelity Images with VQ-VAE-2	arXiv
20/12	VQGAN	Taming Transformers for High-Resolution Image Synthesis	arXiv	GitHub
21/10	ViT-VQGAN	VECTOR-QUANTIZED IMAGE MODELING WITH IMPROVED VQGAN	arXiv
22/02	MaskGIT	MaskGIT: Masked Generative Image Transformer	arXiv
22/09	MoVQ	MoVQ: Modulating Quantized Vectors for High-Fidelity Image Generation	arXiv
22/12	MAGVIT	MAGVIT: Masked Generative Video Transformer	arXiv	GitHub
23/10	Efficient-VQGAN	Efficient-VQGAN: Towards High-Resolution Image Generation with Efficient Vision Transformers	arXiv
24/03	UniCode	UniCode: Learning a Unified Codebook for Multimodal Large Language Models	arXiv
24/05	Chameleon	Chameleon: Mixed-Modal Early-Fusion Foundation Models	arXiv	GitHub
24/05	LG-VQ	LG-VQ: Language-Guided Codebook Learning	arXiv
24/06	LlamaGEN	Autoregressive Model Beats Diffusion: Llama for Scalable Image Generation	arXiv	GitHub
24/06	TiTok	An Image is Worth 32 Tokens for Reconstruction and Generation	arXiv	GitHub
24/06	OmniTokenizer-VQVAE	OmniTokenizer: A Joint Image-Video Tokenizer for Visual Generation	arXiv	GitHub
24/06	VQGAN-LC	Scaling the Codebook Size of VQGAN to 100,000 with a Utilization Rate of 99%	arXiv
24/09	MaskBit	MaskBit: Embedding-free Image Generation via Bit Tokens	arXiv	GitHub
24/10	BPE-VQ	From Pixels to Tokens: Byte-Pair Encoding on Quantized Visual Modalities	arXiv
24/10	RotationTrick	RESTRUCTURING VECTOR QUANTIZATION WITH THE ROTATION TRICK	arXiv	GitHub
24/10	DiGIT	Stabilize the Latent Space for Image Autoregressive Modeling: A Unified Perspective	arXiv	GitHub
24/11	SimVQ	ADDRESSING REPRESENTATION COLLAPSE IN VECTOR QUANTIZED MODELS WITH ONE LINEAR LAYER	arXiv	GitHub
24/11	ALIT	ADAPTIVE LENGTH IMAGE TOKENIZATION VIA RECURRENT ALLOCATION	arXiv	GitHub
24/11	VQ-KD	Image Understanding Makes for A Good Tokenizer for Image Generation	arXiv	GitHub
24/11	FQGAN	Factorized Visual Tokenization and Generation	arXiv	GitHub
24/12	TokenFlow	TokenFlow: Unified Image Tokenizer for Multimodal Understanding and Generation	arXiv	GitHub
24/12	SynerGen-VL	SynerGen-VL: Towards Synergistic Image Understanding and Generation with Vision Experts and Token Folding	arXiv
24/12	SoftVQ-VAE	SoftVQ-VAE: Efficient 1-Dimensional Continuous Tokenizer	arXiv	GitHub
24/12	CRT	When Worse is Better: Navigating the compression-generation tradeoff in visual tokenization	arXiv
24/12	IBQ	Scalable Image Tokenization with Index Backpropagation Quantization	arXiv	GitHub
25/01	TA-TiTok	Democratizing Text-to-Image Masked Generative Models with Compact Text-Aware One-Dimensional Tokens	arXiv	GitHub
25/01	One-D-Piece	One-D-Piece: Image Tokenizer Meets Quality-Controllable Compression	arXiv	GitHub
25/02	UniTok	UniTok: A Unified Tokenizer for Visual Generation and Understanding	arXiv	GitHub
25/03	SemHiTok	SemHiTok: A Unified Image Tokenizer via Semantic-Guided Hierarchical Codebook for Multimodal Understanding and Generation	arXiv
25/03	V2Flow	V2Flow: Unifying Visual Tokenization and Large Language Model Vocabularies for Autoregressive Image Generation	arXiv	GitHub
25/03	Robust Tokenizer	Robust Latent Matters: Boosting Image Generation with Sampling Error	arXiv	GitHub
25/03	PCA Tokenizer	“Principal Components” Enable A New Language of Images	arXiv	GitHub
25/03	DualToken	DualToken: Towards Unifying Visual Understanding and Generation with Dual Visual Vocabularies	arXiv	GitHub
25/03	CTF	Improving Autoregressive Image Generation through Coarse-to-Fine Token Prediction	arXiv	GitHub
25/03	TokenSet	Tokenize Image as a Set	arXiv	GitHub
25/03	TokenBridge	Bridging Continuous and Discrete Tokens for Autoregressive Visual Generation	arXiv	Project Page

Discrete Encoder/RQ

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
22/03	RQ-VAE	Autoregressive Image Generation using Residual Quantization	arXiv	GitHub
24/04	VAR	Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction	arXiv	GitHub
25/03	NFIG	NFIG: Autoregressive Image Generation with Next-Frequency Prediction	arXiv

Discrete Encoder/FSQ

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
23/09	FSQ-VQ-VAE	FINITE SCALAR QUANTIZATION: VQ-VAE MADE SIMPLE	arXiv	GitHub
24/10	ElasticTok	ELASTICTOK: ADAPTIVE TOKENIZATION FOR IMAGE AND VIDEO	arXiv	GitHub
24/12	VIDTOK	VIDTOK: A VERSATILE AND OPEN-SOURCE VIDEO TOKENIZER	arXiv	GitHub
25/02	FlexTok	FlexTok: Resampling Images into 1D Token Sequences of Flexible Length	arXiv

Discrete Encoder/LFQ

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
23/10	MAGVIT-v2	Language Model Beats Diffusion -- Tokenizer is Key to Visual Generation	arXiv
24/05	LIBRA	Libra: Building Decoupled Vision System on Large Language Models	arXiv	GitHub
24/09	Open-MAGVIT2	Open-MAGVIT2: An Open-Source Project Toward Democratizing Auto-regressive Visual Generation	arXiv	GitHub
25/03	FlowMo	Flow to the Mode: Mode-Seeking Diffusion Autoencoders for State-of-the-Art Image Tokenization	arXiv	Project Page

Discrete Encoder/BSQ

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
24/06	BSQ-ViT	Image and Video Tokenization with Binary Spherical Quantization	arXiv	GitHub
25/02	QLIP	QLIP: Text-Aligned Visual Tokenization Unifies Auto-Regressive Multimodal Understanding and Generation	arXiv	GitHub

Discrete Encoder/PQ

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
24/10	ImageFolder	ImageFolder: Autoregressive Image Generation with Folded Tokens	arXiv	GitHub
24/12	XQ-GAN	XQ-GAN: An Open-source Image Tokenization Framework for Autoregressive Generation	arXiv	GitHub

Discrete Encoder/Other Methods

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
24/12	GSQ	Scaling Image Tokenizers with Grouped Spherical Quantization	arXiv	GitHub
24/12	TexTok	Language-Guided Image Tokenization for Generation	arXiv	GitHub
24/12	SIT	Spectral Image Tokenizer	arXiv

Continuous Encoder

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
13/12	VAE	Auto-Encoding Variational Bayes	arXiv
24/12	Divot	Divot: Diffusion Powers Video Tokenizer for Comprehension and Generation	arXiv	GitHub
25/01	VA-VAE	Reconstruction vs. Generation:Taming Optimization Dilemma in Latent Diffusion Models	arXiv	GitHub
25/01	CAT	CAT: Content-Adaptive Image Tokenization	arXiv
25/01	ViTok	Learnings from Scaling Visual Tokenizers for Reconstruction and Generation	arXiv	Project Page
25/02	ReaLS	Exploring Representation-Aligned Latent Space for Better Generation	arXiv	GitHub
25/02	MAETok	Masked Autoencoders Are Effective Tokenizers for Diffusion Models	arXiv	GitHub
25/02	EQ-VAE	EQ-VAE: Equivariance Regularized Latent Space for Improved Generative Image Modeling	arXiv	GitHub
25/03	FAR	Frequency Autoregressive Image Generation with Continuous Tokens	arXiv	GitHub
25/03	USP	USP: Unified Self-Supervised Pretraining for Image Generation and Understanding	arXiv	GitHub
25/03	TokenBridge	Bridging Continuous and Discrete Tokens for Autoregressive Visual Generation	arXiv	Project Page

Text-Representation Encoder

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
23/02	LQAE	Language Quantized AutoEncoders: Towards Unsupervised Text-Image Alignment	arXiv
23/06	SPAE	SPAE: Semantic Pyramid AutoEncoder for Multimodal Generation with Frozen LLMs	arXiv
24/03	V2L-Tokenizer	Beyond Text: Frozen Large Language Models in Visual Signal Comprehension	arXiv	GitHub
24/12	ViLex	Visual Lexicon: Rich Image Features in Language Space	arXiv

External Image Generator

Discrete Condition

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
23/09	LAViT	Unified Language-Vision Pretraining in Llm With Dynamic Discrete Visual Tokenization	arXiv	GitHub
23/10	SEED	Making Llama See and Draw With Seed Tokenizer	arXiv	GitHub
24/02	AnyGPT	Anygpt: Unified Multimodal Llm With Discrete Sequence Modeling	arXiv	GitHub
24/09	MIO	Mio: A Foundation Model on Multimodal Tokens	arXiv	GitHub
24/12	Illume	Illume: Illuminating Your Llms to See, Draw, and Self-Enhance	arXiv	GitHub
25/04	ILLUME+	ILLUME+: Illuminating Unified MLLM with Dual Visual Tokenization and Diffusion Refinement	arXiv	GitHub

Continuous Condition

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
23/07	Emu	Emu: Generative Pretraining in Multimodality	arXiv	GitHub
23/09	NExT-GPT	Next-Gpt	arXiv	GitHub
23/10	MiniGPT-5	Minigpt-5: Interleaved Vision-and-Language Generation Via Generative Vokens	arXiv
23/12	VL-GPT	Vl-Gpt: A Generative Pre-Trained Transformer for Vision and Language Understanding and Generation	arXiv	GitHub
23/12	Emu2	Generative Multimodal Models Are In-Context Learners	arXiv	GitHub
24/01	MM-Interleaved	Mm-Interleaved: Interleaved Image-Text Generative Modeling Via Multi-Modal Feature Synchronizer	arXiv	GitHub
23/10	EasyGen	Easygen: Easing Multimodal Generation With Bidiffuser and Llms	arXiv
23/11	CoDi-2	Codi-2: In-Context Interleaved and Interactive Any-to-Any Generation	arXiv
24/04	SEED-X	Seed-X: Multimodal Models With Unified Multi-Granularity Comprehension and Generation	arXiv	GitHub
23/09	DreamLLM	Dreamllm: Synergistic Multimodal Comprehension and Creation	arXiv	GitHub
24/05	DEEM	Deem: Diffusion Models Serve as the Eyes of Large Language Models for Image Perception	arXiv	GitHub
24/05	X-VILA	X-Vila: Cross-Modality Alignment for Large Language Model	arXiv
24/11	Spider	Spider: Any-to-Many Multimodal Llm	arXiv
24/12	MetaMorph	Metamorph: Multimodal Understanding and Generation Via Instruction Tuning	arXiv	Project Page
25/04	MetaQuery	Transfer between Modalities with MetaQueries	arXiv	Project Page

Discrete Image Modelling

VQGAN Encoder

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
22/02	OFA	Ofa: Unifying Multimodal Pretrained Models	arXiv	GitHub
22/06	Unified-IO	Unified-Io: A Unified Model for Vision, Language, and Multi-Modal Tasks	arXiv	GitHub
23/11	Teal	Teal: Tokenize and Embed All for Multi-Modal Large Language Models	arXiv
24/02	LWM	World Model on Million-Length Video and Language With Ringattention	arXiv	GitHub
24/05	Chameleon	Chameleon: Mixed-Modal Early-Fusion Foundation Models	arXiv	GitHub
24/06	LlamaGEN	Autoregressive Model Beats Diffusion: Llama for Scalable Image Generation	arXiv	GitHub
24/06	4M-21	4M-21: An Any-to-Any Vision Model for Tens of Tasks and Modalities	arXiv
24/07	ANOLE	Anole: An Open, Autoregressive, Native Large Multimodal Models for Interleaved Image-Text Generation	arXiv	GitHub
24/08	Show-o	Show-o: One single transformer to unify multimodal understanding and generation	arXiv
24/09	Emu3	Emu3: Next-Token Prediction is All You Need	arXiv
24/12	Liquid	Liquid: Language Models are Scalable Multi-modal Generators	arXiv
24/12	SynerGen-VL	SynerGen-VL: Towards Synergistic Image Understanding and Generation with Vision Experts and Token Folding	arXiv
25/02	HermesFlow	HermesFlow: Seamlessly Closing the Gap in Multimodal Understanding and Generation	arXiv	GitHub

Semantic Encoders

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
24/05	LIBRA	Libra: Building Decoupled Vision System on Large Language Models	arXiv	GitHub
24/06	SeTok	Towards Semantic Equivalence of Tokenization in Multimodal LLM	arXiv	GitHub
24/09	VILA-U	VILA-U: a Unified Foundation Model Integrating Visual Understanding and Generation	arXiv	GitHub
24/11	MUSE-VL	MUSE-VL: Modeling Unified VLM through Semantic Discrete Encoding	arXiv
24/12	TokenFlow	TokenFlow: Unified Image Tokenizer for Multimodal Understanding and Generation	arXiv	GitHub
25/02	QLIP	QLIP: Text-Aligned Visual Tokenization Unifies Auto-Regressive Multimodal Understanding and Generation	arXiv	GitHub
25/02	UniTok	UniTok: A Unified Tokenizer for Visual Generation and Understanding	arXiv	GitHub
25/03	DualToken	DualToken: Towards Unifying Visual Understanding and Generation with Dual Visual Vocabularies	arXiv	GitHub

Decoupled Encoder

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
23/12	Unified-IO 2	Unified-IO 2: Scaling Autoregressive Multimodal Models with Vision Language Audio and Action	arXiv	GitHub
24/05	Morph-Tokens	Auto-Encoding Morph-Tokens for Multimodal LLM	arXiv	GitHub
24/10	Janus	Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generation	arxiv	GitHub
24/12	ILLUME	ILLUME: Illuminating Your LLMs to See, Draw, and Self-Enhance	arxiv
25/01	VARGPT	VARGPT: Unified Understanding and Generation in a Visual Autoregressive Multimodal Large Language Model	arxiv	GitHub
25/01	Janus-Pro	Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling	arxiv	GitHub
25/03	OmniMamba	OmniMamba: Efficient and Unified Multimodal Understanding and Generation via State Space Models	arxiv	GitHub
25/04	VARGPT-v1.1	VARGPT-v1.1: Improve Visual Autoregressive Large Unified Model via Iterative Instruction Tuning and Reinforcement Learning	arxiv	GitHub
25/04	UniToken	UniToken: Harmonizing Multimodal Understanding and Generation through Unified Visual Encoding	arxiv	GitHub

Continuous Image Modelling

Diffusion

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
24/08	Transfusion	Transfusion: Predict The Next Token and Diffuse Images With One Multi-Modal Model	arXiv
24/09	MonoFormer	MonoFormer: One transformer for both diffusion and autoregression	arXiv	GitHub
24/11	JanusFlow	JanusFlow: Harmonizing Autoregression and Rectified Flow for Unified Multimodal Understanding and Generation	arXiv	GitHub
24/11	JetFormer	JetFormer: An Autoregressive Generative Model of Raw Images and Text	arXiv
24/12	CausalFusion	Causal Diffusion Transformers for Generative Modeling	arXiv	GitHub
24/12	LLaMAFusion	LlamaFusion: Adapting Pretrained Language Models for Multimodal Generation	arXiv

AR+Diffusion

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
24/10	MMAR	Towards Lossless Multi-Modal Auto-Regressive Prababilistic Modeling	arXiv	GitHub
24/12	Orthus	Orthus: Autoregressive Interleaved Image-Text Generation with Modality-Specific Heads	arXiv
24/12	LatentLM	Multimodal Latent Language Modeling with Next-Token Diffusion	arxiv	GitHub
25/03	UniFluid	Unified Autoregressive Visual Generation and Understanding with Continuous Tokens	arxiv

Diffusion Models

Publication Date	Method Abbreviation	Full Title	arXiv Link	Code Repository
23/03	UniDiffuser	One Transformer Fits All Distributions in Multi-Modal Diffusion at Scale	arXiv	GitHub
23/05	CoDi	Any-to-Any Generation via Composable Diffusion	arXiv	Project Page
23/06	UniDiff	UniDiff: Advancing Vision-Language Models with Generative and Discriminative Learning	arXiv
24/12	OmniFlow	OmniFlow: Any-to-Any Generation with Multi-Modal Rectified Flows	arXiv	GitHub
25/01	D-DiT	Dual Diffusion for Unified Image Generation and Understanding	arXiv	GitHub
25/03	X2I	X2I: Seamless Integration of Multimodal Understanding into Diffusion Transformer via Attention Distillation	arXiv	GitHub