| Scaling Properties of Text Conditioning in Visual Generation | ✓ Link | 0.89 | | | | | | | Qwen-Image + Structured Prompter | 2026-07-31 |
| OmniVerifier-M1: Multimodal Meta-Verifier with Explicit Structured Recalibration | | 0.88 | | | | | | | M1-TTS (OmniVerifier-M1 + GPT-Image-1.5) | 2026-05-27 |
| GenRouter: Unified Workflow Routing for Agentic Image Generation | ✓ Link | 0.88 | | | | | | | GenCanvas + GenRouter (Qwen-Image) | 2026-08-17 |
| LVRPO: Language-Visual Alignment with GRPO for Multimodal Understanding and Generation | | 0.85 | 0.86 | 0.78 | 0.93 | 0.89 | 0.87 | 0.79 | LVRPO | 2026-03-29 |
| GenEvolve: Self-Evolving Image Generation Agents via Tool-Orchestrated Visual Experience Distillation | | 0.82 | 0.84 | 0.74 | 0.87 | 0.83 | 0.81 | 0.83 | GenEvolve + Qwen-Image-Edit | 2026-05-20 |
| SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture | ✓ Link | 0.81 | | | | | | | SenseNova-U1-A3B-MoT (w/ CoT) | 2026-05-12 |
| SenseNova-U1.5: Towards Native Unified Visual Intelligence | ✓ Link | 0.81 | 0.82 | | | | | | SenseNova-U1.5-8B-MoT (w/ CoT) | 2026-09-10 |
| Think-Then-Generate: Reasoning-Aware Text-to-Image Diffusion with LLM Encoders | ✓ Link | 0.79 | | | | | | | Qwen-Image (Think-Then-Generate) | 2026-01-15 |
| ToolArtist: Tool-Using Unified Multimodal Models for Agentic Image Generation | | 0.79 | | | | | | | ToolArtist | 2026-08-05 |
| UniReason 1.0: A Unified Reasoning Framework for World Knowledge Aligned Image Generation and Editing | | 0.78 | 0.80 | 0.68 | 0.79 | 0.77 | 0.83 | 0.81 | UniReason 1.0 | 2026-02-02 |
| Mind-Brush: Integrating Agentic Cognitive Search and Reasoning into Image Generation | ✓ Link | 0.78 | | | | | | | Mind-Brush | 2026-02-02 |
| SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture | ✓ Link | 0.78 | 0.78 | 0.73 | 0.82 | 0.80 | 0.85 | 0.77 | SenseNova-U1-SFT (w/ CoT, 8B) | 2026-05-12 |
| Interleaving Reasoning for Better Text-to-Image Generation | ✓ Link | 0.77 | 0.78 | 0.72 | 0.76 | 0.81 | 0.82 | 0.78 | IRG | 2025-09-08 |
| Gen-Searcher: Reinforcing Agentic Search for Image Generation | ✓ Link | 0.77 | | | | | | | Gen-Searcher-8B + Qwen-Image | 2026-03-30 |
| Unify-Agent: A Unified Multimodal Agent for World-Grounded Image Synthesis | | 0.77 | 0.82 | 0.75 | 0.74 | 0.72 | 0.73 | 0.70 | Unify-Agent | 2026-03-31 |
| UniVerse: Empower Unified Generation with Reasoning and Knowledge | | 0.77 | | | | | | | BAGEL w/ CoT (extra data) | 2026-06-03 |
| Think in Strokes, Not Pixels: Process-Driven Image Generation via Interleaved Reasoning | | 0.76 | | | | | | | BAGEL (7B) (process-driven SFT) | 2026-04-08 |
| Unlocking Complex Visual Generation via Closed-Loop Verified Reasoning | | 0.76 | 0.76 | 0.73 | 0.83 | 0.76 | 0.79 | 0.67 | CLVR (9B) | 2026-05-14 |
| Read It Back: Pretrained MLLMs Are Zero-Shot Reward Models for Text-to-Image Generation | ✓ Link | 0.76 | | | | | | | Self-SpectraReward (w/ Self-CoT) | 2026-07-13 |
| Uni-Edit: Intelligent Editing Is A General Task For Unified Model Tuning | ✓ Link | 0.75 | | | | | | | BAGEL-Uni-Edit (Ours) | 2026-05-20 |
| SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture | ✓ Link | 0.74 | 0.73 | 0.69 | 0.80 | 0.73 | 0.82 | 0.69 | SenseNova-U1-SFT (8BA3B) | 2026-05-12 |
| Unlocking Complex Visual Generation via Closed-Loop Verified Reasoning | | 0.74 | 0.73 | 0.72 | 0.83 | 0.73 | 0.80 | 0.64 | CLVR (4B) | 2026-05-14 |
| Learning Native Reflection in Unified Models with Interleaved Reinforcement Learning | ✓ Link | 0.7384 | | | | | | | UMM-Reflection | 2026-09-28 |
| DeepGen 1.0: A Lightweight Unified Multimodal Model for Advancing Image Generation and Editing | | 0.73 | 0.72 | 0.81 | 0.70 | 0.67 | 0.82 | 0.66 | DeepGen 1.0 (RL) | 2026-02-12 |
| InterleaveThinker: Reinforcing Agentic Interleaved Generation | ✓ Link | 0.73 | | | | | | | InterleaveThinker + FLUX.2-klein-9B | 2026-06-11 |
| GenAgent: Scaling Text-to-Image Generation via Agentic Multimodal Reasoning | ✓ Link | 0.72 | 0.78 | 0.67 | 0.78 | 0.72 | 0.71 | 0.55 | GenAgent | 2026-01-26 |
| DeepGen 1.0: A Lightweight Unified Multimodal Model for Advancing Image Generation and Editing | | 0.72 | 0.70 | 0.71 | 0.82 | 0.62 | 0.79 | 0.65 | DeepGen 1.0 (SFT) | 2026-02-12 |
| MindOmni: Unleashing Reasoning Generation in Vision Language Models with RGPO | ✓ Link | 0.71 | 0.75 | 0.70 | 0.76 | 0.76 | 0.72 | 0.52 | MindOmni
(w/ cot) | 2025-05-19 |
| Emerging Properties in Unified Multimodal Pretraining | ✓ Link | 0.70 | 0.76 | 0.69 | 0.75 | 0.65 | 0.75 | 0.58 | Bagel (w/ cot) | 2025-05-20 |
| Understanding-in-Generation: Reinforcing Generative Capability of Unified Model via Infusing Understanding into Generation | ✓ Link | 0.70 | 0.74 | 0.62 | 0.74 | 0.62 | 0.76 | 0.61 | UiG | 2025-09-23 |
| SenseNova-U1.5: Towards Native Unified Visual Intelligence | ✓ Link | 0.7 | | | | | | | SenseNova-U1.5-8B-MoT (w/o thinking) | 2026-09-10 |
| SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture | ✓ Link | 0.69 | 0.65 | 0.66 | 0.82 | 0.68 | 0.81 | 0.66 | SenseNova-U1-SFT (8B) | 2026-05-12 |
| LLaDA2.0-Uni: Unifying Multimodal Understanding and Generation with Diffusion Large Language Model | ✓ Link | 0.68 | 0.54 | 0.77 | 0.82 | 0.79 | 0.87 | 0.60 | LLaDA2.0-Uni | 2026-04-22 |
| STAR: STacked AutoRegressive Scheme for Unified Multimodal Learning | | 0.66 | 0.61 | 0.67 | 0.61 | 0.74 | 0.69 | 0.66 | STAR-7B | 2025-12-15 |
| AutoRubric-T2I: Robust Rule-Based Reward Model for Text-to-Image Alignment | ✓ Link | 0.66 | | | | | | | AutoRubric-T2I (Gemini-3-Flash, trained on PickScore) | 2026-05-20 |
| LongCat-Image Technical Report | ✓ Link | 0.65 | 0.66 | 0.61 | 0.72 | 0.66 | 0.72 | 0.49 | LongCat-Image | 2025-12-08 |
| MUSE: Resolving Manifold Misalignment in Visual Tokenization via Topological Orthogonality | ✓ Link | 0.65 | | | | | | | MUSE-3B | 2026-05-07 |
| EruDiff: Refactoring Knowledge in Diffusion Models for Advanced Text-to-Image Synthesis | ✓ Link | 0.64 | 0.66 | 0.65 | 0.72 | 0.54 | 0.64 | 0.49 | FLUX.1-dev + EruDiff | 2026-03-21 |
| RepFusion: Leveraging Multimodal Priors for Denoising in Representation Space | | 0.64 | | | | | | | RepFusion-SFT w/ Diffusion Decoder | 2026-06-12 |
| SPAR: Semantic-Pixel Self-Alignment and Adaptive Routing for Unified Multimodal Models | | 0.64 | | | | | | | SPAR-3B (extra data) | 2026-06-22 |
| UniLiP: Adapting CLIP for Unified Multimodal Understanding, Generation and Editing | ✓ Link | 0.63 | 0.66 | | | 0.60 | | | UniLIP-3B | 2025-07-31 |
| BLIP3-o: A Family of Fully Open Unified Multimodal Models-Architecture, Training and Dataset | ✓ Link | 0.62 | | | | | | | BLIP3-o (8B) | 2025-05-14 |
| NextFlow: Unified Sequential Modeling Activates Multimodal Understanding and Generation | ✓ Link | 0.62 | | | | | | | NextFlow-RL (7B) | 2026-01-05 |
| From "What" to "How": Constrained Reasoning for Autoregressive Image Generation | | 0.58 | 0.61 | 0.56 | 0.66 | 0.55 | 0.64 | 0.33 | CoR-Painter | 2026-03-03 |
| InternVL-U: Democratizing Unified Multimodal Models for Understanding, Reasoning, Generation and Editing | ✓ Link | 0.58 | | | | | | | InternVL-U (4B, w/ CoT) | 2026-03-10 |
| UniCom: Unified Multimodal Modeling via Compressed Continuous Semantic Representations | | 0.58 | | | | | | | UniCom | 2026-03-11 |
| MUSE: Resolving Manifold Misalignment in Visual Tokenization via Topological Orthogonality | ✓ Link | 0.58 | | | | | | | MUSE-1B | 2026-05-07 |
| LongCat-Next: Lexicalizing Modalities as Discrete Tokens | ✓ Link | 0.57 | | | | | | | LongCat-Next (A3B, 68.5B) | 2026-03-29 |
| UniLiP: Adapting CLIP for Unified Multimodal Understanding, Generation and Editing | ✓ Link | 0.56 | 0.54 | | | 0.50 | | | UniLIP-1B | 2025-07-31 |
| ARM: An AutoRegressive Large Multimodal Model with Unified Discrete Representations | ✓ Link | 0.56 | 0.53 | 0.59 | 0.67 | 0.54 | 0.64 | 0.45 | ARM-RL | 2026-06-09 |
| MANGO: Unleashing Image Generation Capability of Unified Multimodal Models | | 0.56 | | | | | | | MANGO-4B | 2026-09-08 |
| | 0.55 | 0.56 | 0.55 | 0.62 | 0.49 | 0.63 | 0.41 | MetaQuery-XL | |
| UniWorld-V1: High-Resolution Semantic Encoders for Unified Visual Understanding and Generation | ✓ Link | 0.55 | 0.53 | 0.55 | 0.73 | 0.45 | 0.59 | 0.41 | UniWorld-V1 | 2025-06-03 |
| UniCorn: Towards Self-Improving Unified Multimodal Models through Self-Generated Supervision | ✓ Link | 0.55 | | | | | | | UniCorn | 2026-01-06 |
| Rethinking UMM Visual Generation: Masked Modeling for Efficient Image-Only Pre-training | ✓ Link | 0.55 | | | | | | | IOMM-B 512 | 2026-03-17 |
| Intrinsic Distribution Matching | ✓ Link | 0.55 | 0.54 | 0.55 | 0.67 | 0.49 | 0.62 | 0.41 | IDM (Qwen-Image, LoRA r=64, 2 NFE) | 2026-10-29 |
| Self-Adversarial One Step Generation via Condition Shifting | ✓ Link | 0.54 | | | | | | | APEX 20B (SFT) | 2026-04-14 |
| Intrinsic Distribution Matching | ✓ Link | 0.54 | 0.53 | 0.54 | 0.65 | 0.48 | 0.60 | 0.41 | IDM (Qwen-Image, LoRA r=64, 1 NFE) | 2026-10-29 |
| Rethinking UMM Visual Generation: Masked Modeling for Efficient Image-Only Pre-training | ✓ Link | 0.53 | | | | | | | IOMM-L 512 | 2026-03-17 |
| Emerging Properties in Unified Multimodal Pretraining | ✓ Link | 0.52 | 0.44 | 0.55 | 0.68 | 0.44 | 0.60 | 0.39 | Bagel | 2025-05-20 |
| STAR: STacked AutoRegressive Scheme for Unified Multimodal Learning | | 0.52 | 0.58 | 0.54 | 0.48 | 0.49 | 0.51 | 0.54 | STAR-3B | 2025-12-15 |
| Rethinking UMM Visual Generation: Masked Modeling for Efficient Image-Only Pre-training | ✓ Link | 0.50 | | | | | | | IOMM-B 1024 | 2026-03-17 |
| ARM: An AutoRegressive Large Multimodal Model with Unified Discrete Representations | ✓ Link | 0.50 | 0.47 | 0.52 | 0.60 | 0.47 | 0.54 | 0.43 | ARM | 2026-06-09 |
| Playground v2.5: Three Insights towards Enhancing Aesthetic Quality in Text-to-Image Generation | | 0.49 | 0.49 | 0.58 | 0.55 | 0.43 | 0.48 | 0.33 | Playground-v2.5-1024px-aesthetic | 2024-02-27 |
| Rethinking UMM Visual Generation: Masked Modeling for Efficient Image-Only Pre-training | ✓ Link | 0.48 | | | | | | | IOMM-L 1024 | 2026-03-17 |
| PixArt-$α$: Fast Training of Diffusion Transformer for Photorealistic Text-to-Image Synthesis | ✓ Link | 0.47 | 0.45 | 0.50 | 0.48 | 0.49 | 0.56 | 0.34 | PixArt-XL-2-1024-MS | 2023-09-30 |
| Intrinsic Distribution Matching | ✓ Link | 0.47 | 0.46 | 0.47 | 0.58 | 0.42 | 0.54 | 0.37 | IDM (Z-Image, LoRA r=64, 2 NFE) | 2026-10-29 |
| Scaling Rectified Flow Transformers for High-Resolution Image Synthesis | ✓ Link | 0.46 | 0.44 | 0.50 | 0.58 | 0.44 | 0.52 | 0.31 | stable-diffusion-3.5-large | 2024-03-05 |
| Intrinsic Distribution Matching | ✓ Link | 0.46 | 0.45 | 0.46 | 0.56 | 0.41 | 0.52 | 0.37 | IDM (Z-Image, LoRA r=32, 2 NFE) | 2026-10-29 |
| Intrinsic Distribution Matching | ✓ Link | 0.46 | 0.44 | 0.45 | 0.56 | 0.41 | 0.52 | 0.36 | IDM (Z-Image, LoRA r=64, 1 NFE) | 2026-10-29 |
| PairUni: Pairwise Training for Unified Multimodal Language Models | ✓ Link | 0.45 | 0.36 | 0.46 | 0.62 | 0.42 | 0.55 | 0.29 | PairUni-7B | 2025-10-29 |
| SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis | ✓ Link | 0.43 | 0.43 | 0.48 | 0.47 | 0.44 | 0.45 | 0.27 | stable-diffusion-xl-base-0.9 | 2023-07-04 |
| MindOmni: Unleashing Reasoning Generation in Vision Language Models with RGPO | ✓ Link | 0.43 | 0.40 | 0.38 | 0.62 | 0.36 | 0.52 | 0.32 | MindOmni
(w/o cot) | 2025-05-19 |
| LatentUMM: Dual Latent Alignment for Unified Multimodal Models | ✓ Link | 0.418 | | | | | | | Bagel (14B) + LatentUMM | 2026-05-18 |
| Emu3: Next-Token Prediction is All You Need | ✓ Link | 0.39 | 0.34 | 0.45 | 0.48 | 0.41 | 0.45 | 0.27 | Emu3-gen | 2024-09-27 |
| PairUni: Pairwise Training for Unified Multimodal Language Models | ✓ Link | 0.38 | 0.31 | 0.39 | 0.56 | 0.38 | 0.44 | 0.22 | PairUni-1B | 2025-10-29 |
| Show-o: One Single Transformer to Unify Multimodal Understanding and Generation | ✓ Link | 0.35 | 0.28 | 0.40 | 0.48 | 0.30 | 0.46 | 0.30 | Show-o | 2024-08-22 |
| Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling | ✓ Link | 0.35 | 0.30 | 0.37 | 0.49 | 0.36 | 0.42 | 0.26 | Janus-pro | 2025-01-29 |
| Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling | ✓ Link | 0.23 | 0.16 | 0.26 | 0.35 | 0.28 | 0.30 | 0.14 | Janus | 2025-01-29 |