Top Real Time Vision Models

Explore models that run in real-time (or close to real-time).
‍
Deploy select models (i.e. YOLOv8, CLIP) using the Roboflow Hosted API, or your own hardware using Roboflow Inference.

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOv12

Model Size:

Parameters:

Architecture:

YOLO

YOLOv12 is a state-of-the-art computer vision model you can use for detection, segmentation, and more. Learn more »

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

PaliGemma-2

Model Size:

Parameters:

Architecture:

PaliGemma-2 is a multimodal model developed by Google. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLO11

Model Size:

Parameters:

Architecture:

YOLO11 is a computer vision model that you can use for object detection, segmentation, and classification. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

YOLOv9 Image Segmentation

Model Size:

Parameters:

Architecture:

Object Detection

Deploy on Device with Roboflow✅

Object Detection

Florence 2 Image Captioning

Model Size:

Parameters:

Architecture:

Optical Character Recognition

Deploy on Device with Roboflow✅

Optical Character Recognition

Florence 2 OCR

Model Size:

Parameters:

Architecture:

Florence-2 OCR is a subset of Florence-2 that can read characters in images. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

Florence 2 Image Segmentation

Model Size:

Parameters:

Architecture:

Object Detection

Deploy on Device with Roboflow✅

Object Detection

Florence 2 Object Detection

Model Size:

Parameters:

Architecture:

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

Phi-3.5

Model Size:

Parameters:

6600000000.0

Architecture:

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

Segment Anything 2

Model Size:

Parameters:

Architecture:

Segment Anything

Segment Anything 2 (SAM 2) is a real-time image and video segmentation model. Learn more »

Keypoint Detection

Deploy on Device with Roboflow✅

Keypoint Detection

MediaPipe

Model Size:

Parameters:

Architecture:

Object Detection

Deploy on Device with Roboflow✅

Object Detection

RT-DETR

Model Size:

Parameters:

76000000

Architecture:

DETR

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

Cambrian

Model Size:

Parameters:

Architecture:

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

Model Size:

2800000000.0

Parameters:

705000000.0

Architecture:

The 4M model is a versatile multimodal Transformer model developed by EPFL and Apple, capable of handling a handful of vision and language tasks. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

Florence 2

Model Size:

770000000.0

Parameters:

770000000

Architecture:

Florence-2 is a lightweight vision-language model open-sourced by Microsoft under the MIT license. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

PaliGemma Optical Character Recognition

Model Size:

Parameters:

Architecture:

You can use the set of PaliGemma weights trained on the OCRVQA dataset for performing OCR on images. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

PaliGemma Document VQA

Model Size:

Parameters:

Architecture:

You can use the set of PaliGemma weights trained on the DocVQA dataset for asking questions about documents. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

PaliGemma VQA

Model Size:

Parameters:

Architecture:

You can use the set of PaliGemma weights trained on the VQAv2 dataset for asking questions about the contents of images. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

PaliGemma Website Understanding

Model Size:

Parameters:

Architecture:

You can use the set of PaliGemma weights trained on the Screen2Words dataset for asking questions about website screenshots. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

PaliGemma Image Captioning

Model Size:

Parameters:

Architecture:

You can use the set of PaliGemma weights trained on the COCO Captions dataset for zero-shot image captioning. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOv10

Model Size:

Parameters:

29500000

Architecture:

YOLO

YOLOv10 is a real-time object detection model introduced in the paper "YOLOv10: Real-Time End-to-End Object Detection". Learn more »

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

GPT-4o

Model Size:

Parameters:

Architecture:

GPT-4o is OpenAI’s third major iteration of GPT-4 expanding on the capabilities of GPT-4 with Vision Learn more »

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

PaliGemma

Model Size:

4000000000.0

Parameters:

3 Billion

Architecture:

PaliGemma is a vision language model (VLM) by Google that has multimodal capabilities. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

MMOCR

Model Size:

Parameters:

Architecture:

MMOCR is an Optical Character Recognition model zoo implemented with the MMDetection package. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

TrOCR

Model Size:

Parameters:

Architecture:

TrOCR is a Transformer-based OCR model developed by researchers from Microsoft Research. Learn more »

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

Tesseract

Model Size:

Parameters:

Architecture:

Tesseract is a highly popular OCR engine and project, now primarily developed open-source. Learn more »

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

Surya

Model Size:

Parameters:

Architecture:

Surya is a Python package designed for OCR on document layout analysis. Learn more »

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

Google Gemini

Model Size:

Parameters:

Architecture:

Gemini is a family of Large Multimodal Models (LMMs) developed by Google Deepmind focused specifically on multimodality. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

ResNet-50

Model Size:

Parameters:

25600000

Architecture:

Residual Neural Networks

ResNet-50 is a popular image classification model architecture. Learn more »

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

Anthropic Claude 3

Model Size:

Parameters:

Architecture:

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

EasyOCR

Model Size:

Parameters:

50000000

Architecture:

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOv8 Oriented Bounding Boxes

Model Size:

69500000.0

Parameters:

69500000

Architecture:

YOLO

You can retrieve bounding boxes whose edges match an angled object by training an oriented bounding boxes object detection model, such as YOLOv8's Oriented Bounding Boxes model. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

AltCLIP

Model Size:

Parameters:

Architecture:

AltCLIP is a zero-shot image classification model. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

RemoteCLIP

Model Size:

Parameters:

Architecture:

RemoteCLIP is a zero-shot classification model for remote sensing. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

BioCLIP

Model Size:

Parameters:

Architecture:

BioCLIP is a Vision Foundation Model for the Tree of Life Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

MobileCLIP

Model Size:

Parameters:

Architecture:

MobileCLIP is an image embedding model developed by Apple and introduced in the "MobileCLIP: Fast Image-Text Models through Multi-Modal Reinforced Training" paper Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

SigLIP

Model Size:

Parameters:

878000000

Architecture:

SigLIP is an image embedding model defined in the "Sigmoid Loss for Language Image Pre-Training" paper. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOv9

Model Size:

Parameters:

57300000

Architecture:

YOLO

YOLOv9 is an object detection model architecture released on February 21st, 2024. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLO-World

Model Size:

Parameters:

69000000

Architecture:

YOLO

YOLO-World is a zero-shot object detection model. Learn more »

Keypoint Detection

Deploy on Device with Roboflow✅

Keypoint Detection

YOLO-NAS Pose

Model Size:

Parameters:

Architecture:

YOLO

YOLO-NAS Pose is a keypoint detection model developed by Deci AI. Learn more »

Keypoint Detection

Deploy on Device with Roboflow✅

Keypoint Detection

YOLOv8 Pose Estimation

Model Size:

Parameters:

Architecture:

YOLO

The YOLOv8 pose estimation model allows you to detect keypoints in an image. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

Grounded EdgeSAM

Model Size:

Parameters:

Architecture:

Grounded EdgeSAM is a combination of Grounding DINO, a zero-shot object detection model, and EdgeSAM, a fast zero-shot image segmentation model. Learn more »

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

BakLLaVA

Model Size:

Parameters:

13000000000

Architecture:

BakLLaVA is an LMM developed by LAION, Ontocord, and Skunkworks AI. BakLLaVA uses a Mistral 7B base augmented with the LLaVA 1.5 architecture. Learn more »

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

CogVLM

Model Size:

Parameters:

6500000000

Architecture:

CogVLM shows strong performance in Visual Question Answering (VQA) and other vision tasks. Learn more »

Multimodal Model

Deploy on Device with Roboflow✅

Multimodal Model

QwenVL

Model Size:

7000000000.0

Parameters:

Architecture:

Qwen-VL is an LMM developed by Alibaba Cloud. Qwen-VL accepts images, text, and bounding boxes as inputs. The model can output text and bounding boxes. Qwen-VL naturally supports English, Chinese, and multilingual conversation. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

VLPart

Model Size:

Parameters:

Architecture:

VLPart, developed by Meta Research, is an object detection and segmentation model that works with an open vocabulary Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

CoDet

Model Size:

Parameters:

Architecture:

CoDet is an open vocabulary zero-shot object detection model. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

GPT-4 with Vision

Model Size:

Parameters:

Architecture:

Transformer

GPT-4 with Vision is a multimodal language model developed by OpenAI. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

Grounding DINO

Model Size:

Parameters:

Architecture:

Grounding DINO is a state-of-the-art zero-shot object detection model, developed by IDEA Research. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

BLIP

Model Size:

Parameters:

Architecture:

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

Grounded SAM

Model Size:

Parameters:

Architecture:

Combination of Grounding DINO and Segment Anything

GroundedSAM combines Grounding DINO with the Segment Anything Model to identify and segment objects in an image given text captions. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

SAM-CLIP

Model Size:

Parameters:

Architecture:

Combination of Segment Anything and CLIP

Use Grounding DINO, Segment Anything, and CLIP to label objects in images. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

BLIPv2

Model Size:

Parameters:

Architecture:

BLIPv2 is a multimodal model developed by Salesforce Research. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

ALBEF

Model Size:

Parameters:

Architecture:

Object Detection

Deploy on Device with Roboflow✅

Object Detection

OWL ViT

Model Size:

Parameters:

Architecture:

OWL-ViT is a transformer-based object detection model developed by Google Research. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

FastViT

Model Size:

Parameters:

Architecture:

FastViT is a fast image classification model developed by Apple. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

MetaCLIP

Model Size:

Parameters:

Architecture:

CLIP

MetaCLIP is a zero-shot classification and embedding model developed by Meta AI. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

OWLv2

Model Size:

Parameters:

Architecture:

OWLv2 is a transformer-based object detection model developed by Google Research. OWLv2 is the successor to OWL ViT. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

LLaVA-1.5

Model Size:

Parameters:

13000000000

Architecture:

LLaVA is an open source multimodal language model that you can use for visual question answering and has limited support for object detection. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

Kosmos-2

Model Size:

Parameters:

Architecture:

Kosmos-2 is a multimodal language model capable of object detection and grounding text in images. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

L2CS-Net

Model Size:

Parameters:

Architecture:

L2CS-Net is a gaze estimation model that enables you to calculate where someone is looking and in what direction someone is looking. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

DocTR

Model Size:

Parameters:

Architecture:

DocTR is an Optical Character Recognition tool powered by deep learning. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

DINOv2

Model Size:

Parameters:

Architecture:

DINOv2 is a self-supervised method for training computer vision models developed by Meta Research and released in April 2023. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

RTMDet

Model Size:

Parameters:

Architecture:

RTMDet is an efficient real-time object detector, with self-reported metrics outperforming the YOLO series. It achieves 52.8% AP on COCO with 300+ FPS on an NVIDIA 3090 GPU, making it one of the fastest and most accurate object detectors available as of writing this post. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

YOLACT

Model Size:

Parameters:

Architecture:

A simple, fully convolutional model for real-time instance segmentation Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

ByteTrack

Model Size:

Parameters:

Architecture:

ByteTrack is a multi-object tracking computer vision algorithm. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

FastSAM

Model Size:

Parameters:

636000000

Architecture:

FastSAM is an image segmentation model trained using 2% of the data in the Segment Anything Model SA-1B dataset. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

DETIC

Model Size:

25000000.0

Parameters:

Architecture:

Detic is an open source segmentation model developed by Meta Research and released in 2022. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLO-NAS

Model Size:

67000000.0

Parameters:

Architecture:

YOLO

YOLO-NAS is an object detection model developed by Deci that achieves SOTA performances compared to YOLOv5, v7, and v8. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

Segment Anything Model (SAM)

Model Size:

Parameters:

Architecture:

Segment Anything (SAM) is an image segmentation model developed by Meta Research, capable of doing zero-shot segmentation. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

GroundingDINO

Model Size:

Parameters:

Architecture:

Grounding DINO is a zero-shot object detection model made by combining a Transformer-based DINO detector and grounded pre-training. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

DETR

Model Size:

60000000.0

Parameters:

Architecture:

Transformers

Detection Transformer (DETR) is an end-to-end object detection model implemented using the Transformer architecture. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

YOLOv8 Classification

Model Size:

Parameters:

68200000

Architecture:

YOLO

An image classification model built using YOLOv8. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

YOLOv8 Instance Segmentation

Model Size:

Parameters:

68200000

Architecture:

YOLO

The state-of-the-art YOLOv8 model comes with support for instance segmentation tasks. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOv8

Model Size:

68200000.0

Parameters:

Architecture:

YOLO, CNN

YOLOv8 is a state-of-the-art object detection and image segmentation model created by Ultralytics, the developers of YOLOv5. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

YOLOv7 Instance Segmentation

Model Size:

Parameters:

150000000

Architecture:

YOLO

YOLOv7 Instance Segmentation lets you perform segmentation tasks with the YOLOv7 model. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

OneFormer

Model Size:

Parameters:

219 million

Architecture:

Transformers

OneFormer is a state-of-the-art multi-task image segmentation framework that is implemented using transformers. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

ResNet 32

Model Size:

Parameters:

460,000

Architecture:

A fast, simple convolutional neural network that gets the job done for many tasks, including classification. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOX

Model Size:

68.7

Parameters:

99.1 million parameters

Architecture:

CNN, YOLO

YOLOX is a high-performance object detection model. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOR

Model Size:

202.0

Parameters:

12,786,711 (S2D)

Architecture:

CNN, YOLO

YOLOR (You Only Learn One Representation) is an object detection model that uses both implicit and explicit knowledge to make predictions. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOS

Model Size:

Parameters:

85000000

Architecture:

Transformer, YOLO

YOLOS looks at patches of an image to to form "patch tokens", which are used in place of the traditional wordpiece tokens in NLP. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

Scaled YOLOv4

Model Size:

Parameters:

Architecture:

YOLO

Scaled YOLOv4 is an extension of the YOLOv4 research implemented in the YOLOv5 PyTorch framework. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

Vision Transformer

Model Size:

Parameters:

Architecture:

The Vision Transformer leverages powerful natural language processing embeddings (BERT) and applies them to images. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

Mask RCNN

Model Size:

Parameters:

Architecture:

Mask RCNN is a convolutional neural network for instance segmentation. Learn more »

Semantic Segmentation

Deploy on Device with Roboflow✅

Semantic Segmentation

SegFormer

Model Size:

Parameters:

Architecture:

Transformers

SegFormer is a computer vision framework used in semantic segmentation tasks, implemented with transformers. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

YOLOv5 Classification

Model Size:

Parameters:

Architecture:

YOLO

YOLOv5 Classification is a version of the YOLOv5 model used in single-label and multi-label image classification. Learn more »

Instance Segmentation

Deploy on Device with Roboflow✅

Instance Segmentation

YOLOv5 Instance Segmentation

Model Size:

Parameters:

Architecture:

CNN, YOLO

YOLOv5 Instance Segmentation is a version of YOLOv5 that can be used for instance segmentation tasks. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOv7

Model Size:

75.6

Parameters:

Architecture:

YOLO, CNN

YOLOv7 is a state of the art object detection model. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

MT-YOLOv6

Model Size:

Parameters:

Architecture:

CNN, YOLO

MT-YOLOv6 is a YOLO based model released in 2022. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOv5 Oriented Bounding Boxes

Model Size:

Parameters:

7.5 million

Architecture:

CNN, YOLO

YOLOv5-OBB is a variant of YOLOv5 that supports oriented bounding boxes. This model is designed to yield predictions that better fit objects that are positioned at an angle. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

OpenAI CLIP

Model Size:

Parameters:

428000000.0

Architecture:

CLIP (Contrastive Language-Image Pre-Training) is an impressive multimodal zero-shot image classifier that achieves impressive results in a wide range of domains with no fine-tuning. It applies the recent advancements in large-scale transformers like GPT-3 to the vision arena. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

ResNet 34

Model Size:

Parameters:

Architecture:

A fast, simple convolutional neural network that gets the job done for many tasks, including classification. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

MobileNet V2 Classification

Model Size:

Parameters:

Architecture:

MobileNet is a GoogleAI model well-suited for on-device, real-time classification (distinct from MobileNetSSD, Single Shot Detector). This implementation leverages transfer learning from ImageNet to your dataset. Learn more »

Classification

Deploy on Device with Roboflow✅

Classification

EfficientNet

Model Size:

Parameters:

Architecture:

CNN

EfficientNet is from a family of image classification models from GoogleAI that train comparatively quickly on small amounts of data, making the most of limited datasets. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOv4 PyTorch

Model Size:

Parameters:

Architecture:

YOLO

YOLOv4 has emerged as the best real time object detection model. YOLOv4 carries forward many of the research contributions of the YOLO family of models along with new modeling and data augmentation techniques. This implementation is in PyTorch. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

MobileNet SSD v2

Model Size:

Parameters:

Architecture:

This architecture provides good realtime results on limited compute. It's designed to run in realtime (30 frames per second) even on mobile devices. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOv3 Keras

Model Size:

Parameters:

Architecture:

YOLO

Though it is no longer the most accurate object detection algorithm, YOLO v3 is still a very good choice when you need real-time detection while maintaining excellent accuracy. Keras implementation. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

YOLOv3 PyTorch

Model Size:

Parameters:

Architecture:

YOLO

Though it is no longer the most accurate object detection algorithm, YOLO v3 is still a very good choice when you need real-time detection while maintaining excellent accuracy. PyTorch version. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

Faster R-CNN

Model Size:

Parameters:

Architecture:

One of the most accurate object detection algorithms but requires a lot of power at inference time. A good choice if you can do processing asynchronously on a server. Learn more »

Object Detection

Deploy on Device with Roboflow✅

Object Detection

EfficientDet

Model Size:

Parameters:

3.9 million

Architecture:

EfficientDet achieves the best performance in the fewest training epochs among object detection model architectures, making it a highly scalable architecture especially when operating with limited compute. Learn more »

Deploy a computer vision model today

Join 800,000+ developers curating high quality datasets and deploying better models with Roboflow.

Get started

Explore Models

By Task

By Feature

Top Real Time Vision Models

Deploy a computer vision model today