NEWAgents can now see video via MCP.Try it now →
    Models/Detection & Recognition/ultralytics/yolov8n
    PyTorchObject DetectionAGPL-3.0

    yolov8n

    by ultralytics

    State-of-the-art real-time object detection, YOLO v8 Nano

    dl/month
    3.2Mparams
    Identifiers
    Model ID
    ultralytics/yolov8n
    Feature URI
    mixpeek://image_extractor@v1/ultralytics_yolov8n_v1

    Overview

    YOLOv8n is the nano variant of Ultralytics' YOLOv8 family, optimized for real-time inference while maintaining strong detection accuracy. It supports detection, segmentation, classification, and pose estimation tasks.

    On Mixpeek, YOLOv8n is the fastest object detection option, ideal for processing large video libraries where throughput matters more than detecting every small object.

    Architecture

    CSPDarknet53 backbone with C2f modules, FPN+PAN neck, and decoupled head. Anchor-free detection with distribution focal loss. Nano variant uses depth=0.33, width=0.25.

    Mixpeek SDK Integration

    import { Mixpeek } from "mixpeek";
    const mx = new Mixpeek({ apiKey: "API_KEY" });
    await mx.collections.ingest({
    collection_id: "my-collection",
    source: { url: "https://example.com/video.mp4" },
    feature_extractors: [{
    name: "object_detection",
    version: "v1",
    params: {
    model_id: "ultralytics/yolov8n"
    }
    }]
    });

    Capabilities

    • Real-time inference at 640px resolution
    • 80 COCO object categories
    • Multi-task: detection, segmentation, classification, pose
    • Optimized for edge and mobile deployment

    Use Cases on Mixpeek

    Live video stream monitoring with sub-10ms latency
    Sports analytics, track players, balls, equipment in real-time
    Manufacturing QA, detect defects on production lines

    Benchmarks

    DatasetMetricScoreSource
    COCO val2017mAP50-9537.3Ultralytics YOLOv8 docs
    COCO val2017mAP5052.6Ultralytics YOLOv8 docs

    Performance

    Input Size640×640 px
    GPU Latency~1.2ms / image (A100)
    CPU Latency~18ms / image
    GPU Throughput~830 images/sec (A100)
    GPU Memory~0.3 GB

    3.2M params — nano variant for real-time detection

    Specification

    FrameworkPyTorch
    Organizationultralytics
    FeatureObject Detection
    Outputbbox + label
    Modalitiesvideo, image
    RetrieverObject Filter
    Parameters3.2M
    LicenseAGPL-3.0
    Downloads/mo

    Build a pipeline with yolov8n

    Add this model to a processing pipeline alongside other extractors. Combine with retrieval stages for end-to-end search.

    Open Pipeline Builder