mlsanityData to 14 Oct 2026

Emerging Properties in Unified Multimodal Pretraining

arXiv 2505.14683 ↗First posted 20 May 2025cs.CVcited by 897

Not on the map. Why?

Cited by 160 in LLM reasoning, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Unifying multimodal understanding and generation has shown impressive capabilities in cutting-edge proprietary systems. In this work, we introduce BAGEL, an open-source foundational model that natively supports multimodal understanding and generation. BAGEL is a unified, decoder-only model pretrained on trillions of tokens curated from large-scale interleaved text, image, video, and web data. When scaled with such diverse multimodal interleaved data, BAGEL exhibits emerging capabilities in complex multimodal reasoning. As a result, it significantly outperforms open-source unified models in both multimodal generation and understanding across standard benchmarks, while exhibiting advanced multimodal reasoning abilities such as free-form image manipulation, future frame prediction, 3D manipulation, and world navigation. In the hope of facilitating further opportunities for multimodal research, we share the key findings, pretraining details, data creation protocal, and release our code and checkpoints to the community. The project page is at https://bagel-ai.org/

Key facts. Cited by 897 papers in the map (data through 14 Oct 2026). Builds on 87 of its 101 references that are in the map. First posted on arXiv 20 May 2025.

Builds on

87 of 101 references.

References · 87

  1. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  2. 2022Training language models to follow instructions with human feedbackcited by 15,971
  3. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  4. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
  5. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,585
  6. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,355
  7. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,330
  8. 2022Scalable Diffusion Models with Transformerscited by 6,054
  9. 2022Classifier-Free Diffusion Guidancecited by 5,847
  10. 2022Flow Matching for Generative Modelingcited by 5,702
  11. 2025Qwen2.5-VL Technical Reportcited by 5,587
  12. 2024Qwen2.5 Technical Reportcited by 4,942
38 more
  1. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,703
  2. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,356
  3. 2024Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolutioncited by 4,234
  4. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,181
  5. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,700
  6. 2024LLaVA-OneVision: Easy Visual Task Transfercited by 2,719
  7. 2022InstructPix2Pix: Learning to Follow Image Editing Instructionscited by 2,611
  8. 2022Emergent Abilities of Large Language Modelscited by 2,512
  9. 2023MMBench: Is Your Multi-modal Model an All-around Player?cited by 2,141
  10. 2023MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGIcited by 2,080
  11. 2020GLU Variants Improve Transformercited by 1,804
  12. 2023MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Modelscited by 1,652
  13. 2023MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contextscited by 1,614
  14. 2024Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scalingcited by 1,599
  15. 2019Root Mean Square Layer Normalizationcited by 1,480
  16. 2023GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpointscited by 1,346
  17. 2023MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilitiescited by 1,239
  18. 2025SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Featurescited by 1,188
  19. 2024How Far Are We to GPT-4V? Closing the Gap to Commercial Multimodal Models with Open-Source Suitescited by 1,050
  20. 2023GenEval: An Object-Focused Framework for Evaluating Text-to-Image Alignmentcited by 972
  21. 2024Chameleon: Mixed-Modal Early-Fusion Foundation Modelscited by 927
  22. 2025Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scalingcited by 812
  23. 2024Emu3: Next-Token Prediction is All You Needcited by 781
  24. 2024Show-o: One Single Transformer to Unify Multimodal Understanding and Generationcited by 752
  25. 2024Eyes Wide Shut? Exploring the Visual Shortcomings of Multimodal LLMscited by 735
  26. 2024MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategiescited by 723
  27. 2022Autoregressive Image Generation using Residual Quantizationcited by 722
  28. 2023Scaling Vision Transformers to 22 Billion Parameterscited by 694
  29. 2024Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusioncited by 681
  30. 2021Vector-quantized Image Modeling with Improved VQGANcited by 658
  31. 2023MagicBrush: A Manually Annotated Dataset for Instruction-Guided Image Editingcited by 624
  32. 2024DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understandingcited by 572
  33. 2024Transfusion: Predict the Next Token and Diffuse Images with One Multi-Modal Modelcited by 483
  34. 2024Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generationcited by 457
  35. 2023Generative Multimodal Models are In-Context Learnerscited by 446
  36. 2023Finite Scalar Quantization: VQ-VAE Made Simplecited by 440
  37. 2025Step1X-Edit: A Practical Framework for General Image Editingcited by 429
  38. 2024OmniGen: Unified Image Generationcited by 394

50 most cited of 87 All 87 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 286; 2026: 611.

897 papers.

Sort:

By area:

  1. 2025Qwen-Image Technical Reportcited by 1,044
  2. 2025FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Spacecited by 1,006
  3. 2026World Action Models are Zero-shot Policiescited by 465
  4. 2026Causal World Modeling for Robot Controlcited by 321
  5. 2025OmniGen2: Towards Instruction-Aligned Multimodal Generationcited by 253
  6. 2025Motus: A Unified Latent Action World Modelcited by 251
  7. 2025Show-o2: Improved Native Unified Multimodal Modelscited by 250
  8. 2025SAM 3D: 3Dfy Anything in Imagescited by 247
  9. 2025Thinking with Images for Multimodal Reasoning: Foundations, Methods, and Future Frontierscited by 199
  10. 2026Cosmos 3: Omnimodal World Models for Physical AIcited by 165
  11. 2026Seedance 2.0: Advancing Video Generation for World Complexitycited by 160
  12. 2025HunyuanImage 3.0 Technical Reportcited by 149
38 more
  1. 2026π0.7: a Steerable Generalist Robotic Foundation Model with Emergent Capabilitiescited by 138
  2. 2025X-Omni: Reinforcement Learning Makes Discrete Autoregressive Image Generative Models Great Againcited by 112
  3. 2025Latent Visual Reasoningcited by 104
  4. 2026A Pragmatic VLA Foundation Modelcited by 103
  5. 2025Echo-4o: Harnessing the Power of GPT-4o Synthetic Images for Improved Image Generationcited by 100
  6. 2025LongCat-Image Technical Reportcited by 100
  7. 2025Pref-GRPO: Pairwise Preference Reward-based GRPO for Stable Text-to-Image Reinforcement Learningcited by 92
  8. 2025Lumina-DiMOO: An Omni Diffusion Large Language Model for Multi-Modal Generation and Understandingcited by 88
  9. 2025OneIG-Bench: Omni-dimensional Nuanced Evaluation for Image Generationcited by 87
  10. 2025UniVideo: Unified Understanding, Generation, and Editing for Videoscited by 87
  11. 2025F1: A Vision-Language-Action Model Bridging Understanding and Generation to Actionscited by 86
  12. 2025Zebra-CoT: A Dataset for Interleaved Vision Language Reasoningcited by 84
  13. 2025Scaling Spatial Intelligence with Multimodal Foundation Modelscited by 82
  14. 2025Uniworld-V2: Reinforce Image Editing with Diffusion Negative-aware Finetuning and MLLM Implicit Feedbackcited by 79
  15. 2025InstructVLA: Vision-Language-Action Instruction Tuning from Understanding to Manipulationcited by 72
  16. 2025KRIS-Bench: Benchmarking Next-Level Intelligent Image Editing Modelscited by 71
  17. 2025Chain-of-Visual-Thought: Teaching VLMs to See and Think Better with Continuous Visual Tokenscited by 71
  18. 2025TIIF-Bench: How Does Your T2I Model Follow Your Instructions?cited by 69
  19. 2025Ovis-U1 Technical Reportcited by 67
  20. 2025UniWorld-V1: High-Resolution Semantic Encoders for Unified Visual Understanding and Generationcited by 66
  21. 2026Being-H0.5: Scaling Human-Centric Robot Learning for Cross-Embodiment Generalizationcited by 66
  22. 2025EditScore: Unlocking Online RL for Image Editing via High-Fidelity Reward Modelingcited by 64
  23. 2025EditReward: A Human-Aligned Reward Model for Instruction-Guided Image Editingcited by 62
  24. 2025NextStep-1: Toward Autoregressive Image Generation with Continuous Tokens at Scalecited by 57
  25. 2025Uni-cot: Towards Unified Chain-of-Thought Reasoning Across Text and Visioncited by 56
  26. 2025ThinkMorph: Emergent Properties in Multimodal Interleaved Chain-of-Thought Reasoningcited by 56
  27. 2025EditVerse: Unifying Image and Video Editing and Generation with In-Context Learningcited by 54
  28. 2025BLIP3o-NEXT: Next Frontier of Native Image Generationcited by 54
  29. 2026Scaling Text-to-Image Diffusion Transformers with Representation Autoencoderscited by 53
  30. 2025GenEval 2: Addressing Benchmark Drift in Text-to-Image Evaluationcited by 52
  31. 2025OneCAT: Decoder-Only Auto-Regressive Model for Unified Understanding and Generationcited by 50
  32. 2026InternVLA-A1: Unifying Understanding, Generation and Action for Robotic Manipulationcited by 50
  33. 2026Zooming without Zooming: Region-to-Image Distillation for Fine-Grained Multimodal Perceptioncited by 50
  34. 2026World Action Models: The Next Frontier in Embodied AIcited by 50
  35. 2025Reconstruction Alignment Improves Unified Multimodal Modelscited by 49
  36. 2025FUDOKI: Discrete Flow-based Unified Understanding and Generation via Kinetic-Optimal Velocitiescited by 46
  37. 2025OpenUni: A Simple Baseline for Unified Multimodal Understanding and Generationcited by 46
  38. 2025UniLiP: Adapting CLIP for Unified Multimodal Understanding, Generation and Editingcited by 44

50 most cited of 897 All 897 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025MME-Unify: A Comprehensive Benchmark for Unified Multimodal Understanding and Generation Models
  2. 2026UniM: A Unified Any-to-Any Interleaved Multimodal Benchmark
  3. 2025DiffThinker: Towards Generative Multimodal Reasoning with Diffusion Models
  4. 2025R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalization
  5. 2026WorldBench: A Challenging and Visually Diverse Multimodal Reasoning Benchmark
  6. 2025OmniMamba: Efficient and Unified Multimodal Understanding and Generation via State Space Models
  7. 2025Beyond Seeing: Evaluating Multimodal LLMs on Tool-Enabled Image Perception, Transformation, and Reasoning
  8. 2025LENS: Multi-level Evaluation of Multimodal Reasoning with Large Language Models