Model comparison · YOLOv10 vs YOLOv9

YOLOv10 vs YOLOv9

A size-tier-by-size-tier comparison of YOLOv10 and YOLOv9: published COCO accuracy, parameter counts, licensing and which one to reach for. Every figure is the model authors' own, linked to its source.

CompareYOLOv10
AgainstYOLOv9
49%fewer params

mAP comparison

YOLOv10YOLOv9
35.541.347.052.858.5mAPNanoSmallMediumLargeExtra largeModel size tierYOLOv10n: 38.5 mAP, 2.3M paramsYOLOv10s: 46.3 mAP, 7.2M paramsYOLOv10m: 51.1 mAP, 15.4M paramsYOLOv10l: 53.2 mAP, 24.4M paramsYOLOv10x: 54.4 mAP, 29.5M paramsYOLOv9t: 38.3 mAP, 2M paramsYOLOv9s: 46.8 mAP, 7.2M paramsYOLOv9m: 51.4 mAP, 20.1M paramsYOLOv9c: 53.0 mAP, 25.5M paramsYOLOv9e: 55.6 mAP, 58.1M params

YOLOv10

  • Nano 38.5
  • Small 46.3
  • Medium 51.1
  • Large 53.2
  • Extra large 54.4

YOLOv9

  • Nano 38.3
  • Small 46.8
  • Medium 51.4
  • Large 53.0
  • Extra large 55.6
COCO mAP at 640px, as published by each model’s authors. The y axis spans 35.5 to 58.5 rather than starting at zero, so small differences stay visible. Points are the published checkpoints; the joining lines are for reading order only.

Accuracy and size, tier by tier

Each row pairs the equivalent rung of both size ladders, so the numbers sit side by side at comparable capacity. Every figure is the model authors’ own.

  • COCO mAP50-95 @ 640px
  • mAP: higher is better
  • Params (M): lower is better
YOLOv10 and YOLOv9 compared at each size tier: published COCO mAP50-95 at 640px and parameter counts in millions.
TierYOLOv10YOLOv9Δ mAP
ModelmAPParamsModelmAPParams
NanoYOLOv10n38.52.3MYOLOv9t38.32.0M+0.2
SmallYOLOv10s46.37.2MYOLOv9s46.87.2M-0.5
MediumYOLOv10m51.115.4MYOLOv9m51.420.1M-0.3
LargeYOLOv10l53.224.4MYOLOv9c53.025.5M+0.2
Extra largeYOLOv10x54.429.5MYOLOv9e55.658.1M-1.2

Δ mAP is YOLOv10 minus YOLOv9: positive means YOLOv10 is ahead at that tier. Sources: YOLOv10 docs · YOLOv10 paper · YOLOv9 docs · YOLOv9 paper

YOLOv10 also ships YOLOv10b, which sits between the rungs above and has no direct counterpart in YOLOv9, so it is left out of the paired rows rather than matched to an unequal one.

Additional information

  • YOLOv9 posts the higher published COCO mAP at 3 of the 5 matched size tiers, averaging 0.32 mAP ahead of YOLOv10. The gap is widest at the extra large tier, where YOLOv9 leads by 1.2 mAP. At the top of each range, YOLOv9 reaches 55.6 mAP against 54.4.

  • 49% fewer parameters at the extra large tier. YOLOv10x reaches 54.4 mAP from 29.5M parameters against YOLOv9e at 55.6 mAP from 58.1M, so the saving costs 1.2 mAP. No checkpoint from either family matches a larger one from the other outright.

    Parameter count is not latency. It is a good proxy for download size and memory, but throughput depends on your hardware, batch size and export format, so measure on your own target.

YOLOv10

The first popular NMS-free, end-to-end YOLO.

Released
2024 · Tsinghua University (THU-MIG)
License
AGPL-3.0
Framework
Built on Ultralytics (PyTorch)
Tasks
Detect

The first widely used YOLO that skips the NMS post-processing step, using paired training heads so inference is truly end to end and low latency.

Download YOLOv10 weights →

YOLOv9

Fixes deep-network information loss with PGI and GELAN.

Released
2024 · Academia Sinica (Wang et al.)
License
GPL-3.0
Framework
Original repo + Ultralytics (PyTorch)
Tasks
Detect, Segment

Adds Programmable Gradient Information and the GELAN backbone to preserve detail through deep layers, reaching high accuracy with very few parameters.

Download YOLOv9 weights →

YOLOv10 and YOLOv9 against the whole lineage

YOLOv10 and YOLOv9 against every other detector here, on one shared scale. The rest are dimmed; hover any row to bring it back.

Every YOLO detector in this section, with release year, publisher, published COCO mAP50-95 range at 640px, parameter-count range and licence.
ModelReleasedmAP rangeParams rangeLicence
YOLO26202640.9 to 57.52.4M to 55.7MAGPL-3.0
YOLO12202540.6 to 55.22.6M to 59.1MAGPL-3.0
YOLO11202439.5 to 54.72.6M to 56.9MAGPL-3.0
YOLOv10202438.5 to 54.42.3M to 29.5MAGPL-3.0
YOLOv9202438.3 to 55.62.0M to 58.1MGPL-3.0
YOLOv8202337.3 to 53.93.2M to 68.2MAGPL-3.0
YOLOv5 (v5u)202034.3 to 53.22.6M to 97.2MAGPL-3.0

Licenses here cover the weights. Running either model through the Ultralytics package brings AGPL-3.0 with it, and that can reach your own application regardless of how the checkpoint itself is licensed. Check the terms against your distribution plan before you ship.