mlsanityData to 14 Oct 2026

SAM 3: Segment Anything with Concepts

arXiv 2511.16719 ↗First posted 20 Nov 2025cs.CV cs.AIcited by 906

Not on the map. Why?

Cited by 96 in Multimodal LLMs / vision-language models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

We present Segment Anything Model (SAM) 3, a unified model that detects, segments, and tracks objects in images and videos based on concept prompts, which we define as either short noun phrases (e.g., "yellow school bus"), image exemplars, or a combination of both. Promptable Concept Segmentation (PCS) takes such prompts and returns segmentation masks and unique identities for all matching object instances. To advance PCS, we build a scalable data engine that produces a high-quality dataset with 4M unique concept labels, including hard negatives, across images and videos. Our model consists of an image-level detector and a memory-based video tracker that share a single backbone. Recognition and localization are decoupled with a presence head, which boosts detection accuracy. SAM 3 doubles the accuracy of existing systems in both image and video PCS, and improves previous SAM capabilities on visual segmentation tasks. We open source SAM 3 along with our new Segment Anything with Concepts (SA-Co) benchmark for promptable concept segmentation.

Key facts. Cited by 906 papers in the map (data through 14 Oct 2026). First posted on arXiv 20 Nov 2025.

Builds on

0 of 0 references.

No references here yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 25; 2026: 881.

906 papers.

Sort:

By area:

  1. 2026Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Groundingcited by 139
  2. 2026CaP-X: A Framework for Benchmarking and Improving Coding Agents for Robot Manipulationcited by 88
  3. 2026WorldArena: A Unified Benchmark for Evaluating Perception and Functional Utility of Embodied World Modelscited by 76
  4. 2026Zooming without Zooming: Region-to-Image Distillation for Fine-Grained Multimodal Perceptioncited by 50
  5. 2026SceneSmith: Agentic Generation of Simulation-Ready Indoor Scenescited by 46
  6. 2026HY-World 2.0: A Multi-Modal World Model for Reconstructing, Generating, and Simulating 3D Worldscited by 44
  7. 2026Image Generators are Generalist Vision Learnerscited by 43
  8. 2026FireRed-Image-Edit-1.0 Technical Reportcited by 36
  9. 2026Kiwi-Edit: Versatile Video Editing via Instruction and Reference Guidancecited by 36
  10. 2026LLM-as-a-Verifier: A General-Purpose Verification Frameworkcited by 30
  11. 2026VGGT-SLAM 2.0: Real-time Dense Feed-forward Scene Reconstructioncited by 28
  12. 2026LiveWorld: Simulating Out-of-Sight Dynamics in Generative Video World Modelscited by 28
38 more
  1. 2026OA-WAM: Object-Addressable World Action Model for Robust Robot Manipulationcited by 28
  2. 2026OmniWeaving: Towards Unified Video Generation with Free-form Composition and Reasoningcited by 27
  3. 2025MedSAM3: Delving into Segment Anything with Medical Conceptscited by 26
  4. 2026Medical SAM3: A Foundation Model for Universal Prompt-Driven Medical Image Segmentationcited by 25
  5. 2026Latent Reasoning VLA: Latent Thinking and Prediction for Vision-Language-Action Modelscited by 22
  6. 2026RHO: Your Coding Agent is Secretly a Roboticistcited by 20
  7. 2026VP-VLA: Visual Prompting as an Interface for Vision-Language-Action Modelscited by 17
  8. 2026SpatialClaw: Rethinking Action Interface for Agentic Spatial Reasoningcited by 17
  9. 2025SegEarth-OV3: Exploring SAM 3 for Open-Vocabulary Semantic Segmentation in Remote Sensing Imagescited by 16
  10. 2026MaskWAM: Unifying Mask Prompting and Prediction for World-Action Modelscited by 16
  11. 2026FlowWAM: Optical Flow as a Unified Action Representation for World Action Modelscited by 16
  12. 2026MolmoPoint: Better Pointing for VLMs with Grounding Tokenscited by 15
  13. 2026EponaV2: Driving World Model with Comprehensive Future Reasoningcited by 15
  14. 2026Walk through Paintings: Egocentric World Models from Internet Priorscited by 14
  15. 2026EgoSim: Egocentric World Simulator for Embodied Interaction Generationcited by 14
  16. 2026Video Generation Models are General-Purpose Vision Learnerscited by 14
  17. 2026Code as Worlds: Agentic Discovery of Executable World Representations for Physical Reasoningcited by 14
  18. 2026The Second Challenge on Cross-Domain Few-Shot Object Detection at NTIRE 2026: Methods and Resultscited by 13
  19. 2026Grounding World Simulation Models in a Real-World Metropoliscited by 12
  20. 2026INSID3: Training-Free In-Context Segmentation with DINOv3cited by 12
  21. 2026JoyAI-Image: Awaking Spatial Intelligence in Unified Multimodal Understanding and Generationcited by 12
  22. 2026MemoBench: Benchmarking World Modeling in Dynamically Changing Environmentscited by 12
  23. 2026Articulated 3D Scene Graphs for Open-World Mobile Manipulationcited by 11
  24. 2026RoboStereo: Dual-Tower 4D Embodied World Models for Unified Policy Optimizationcited by 10
  25. 2026SegviGen: Repurposing 3D Generative Model for Part Segmentationcited by 10
  26. 2026Efficient Universal Perception Encodercited by 10
  27. 2026SpatialEdit: Benchmarking Fine-Grained Image Spatial Editingcited by 10
  28. 2026ACE-Ego-0: Unifying Egocentric Human and Robotic Data for VLA Pretrainingcited by 10
  29. 2026Guava: Distilling Frontier VLMs into a Compact Agent through a Robotic Manipulation Harnesscited by 10
  30. 2025SAM3-I: Segment Anything with Instructionscited by 9
  31. 2026SAW: Toward a Surgical Action World Model via Controllable and Scalable Video Generationcited by 9
  32. 2026VOID: Video Object and Interaction Deletioncited by 9
  33. 2026Masked Visual Actions for Unified World Modelingcited by 9
  34. 2026TC-IDM: Grounding Video Generation for Executable Zero-shot Robot Motioncited by 8
  35. 2026NovaPlan: Zero-Shot Long-Horizon Manipulation via Closed-Loop Video Language Planningcited by 8
  36. 2026Gaze2Act: Gaze-Conditioned Vision-Language-Action Policies for Interactive Robot Manipulationcited by 8
  37. 2026PAI-Studio: Cinematic Video Background Replacement with Camera-Aware Motioncited by 8
  38. 2026SCAIL-2: Unifying Controlled Character Animation with End-to-End In-Context Conditioningcited by 8

50 most cited of 906 All 906 (JSON)