Nvidia RTX 3060 · rf-detr

Frigate 0.16Community build

Posted Aug 25, 2025·GitHub #19741 ·db-139

I run this too, add my numbers Opens the submit form prefilled with this hardware, model and version

Inference

28–29 ms

as reported

Est. capacity

~35 det/s

1000 ÷ 28.5 ms · estimate

Cameras

—

not reported

CPU

—

not reported

Power

—

not reported

Parts

prices show only once verified by hand

Accelerator

Nvidia RTX 3060

ONNX (Nvidia / GPU), TensorRT (Nvidia, legacy) · 4 builds on DetectFPS · reported as “Nvidia RTX 3060 12GB”

price not tracked yet

Get price drops in the weekly newsletter

Price drops, once a week

When a tracked price is verified or drops, it goes in the newsletter alongside new benchmarks. No per-device alerts and no tracking: the newsletter is the only list.

Newsletter coming soon The signup form appears here as soon as the newsletter embed code is pasted into site.json. Issue #1 is already readable here.

Host

—

not reported

—

CPU

—

not reported

—

Cameras

—

resolution not reported

—

Tracked totalno verified prices yet

Setup

Detector

ONNX (Nvidia / GPU)

Detector instances

—

Model, as reported

rf-detr

Model family · input

rf-detr · input not stated

Camera resolution

—

Cameras

—

Frigate version

0.16 (0.16 series, 2 behind newest)

—

Numbers

dashes mean the source didn't say

Inference

—

Inference range

28–29 ms

Total fps

—

CPU

—

Power

—

Used in stats

No (left out of medians, leaderboards and the calculator)

Notes and flags

Problem report. Low GPU utilization issue.

(low GPU utilization issue)

Source

Where

github.com/blakeblackshear/frigate/discussions/19741

Date, as the source wrote it

2025-08-25 (day precision)

Added to DetectFPS

from a public post; numbers copied as written, not re-tested

What this build implies

~35 detections per second

1000 ÷ 28.5 ms. No camera count was reported, so use the calculator to see what that covers: at 5 fps and Typical activity, that is about 8 cameras comfortably.

Open in the calculator

Hardware

Nvidia RTX 3060

Detector type

ONNX (Nvidia / GPU), TensorRT (Nvidia, legacy)

Builds on DetectFPS

Price

price not tracked yet

Shown because this guide lists Nvidia RTX 3060 in its hardware (hardware_ids). Cross-device guides are only linked when they name this device or its host.

Affiliate links

DetectFPS may earn a commission when you buy through links on this site. It never changes which builds we show or how they rank. Buy buttons say “affiliate link” whenever one is used; none is used on this page.

Similar builds

Same device, then the same model family

Nvidia RTX 3060

TensorRT (Nvidia, legacy)

older version, may be outdated

Speed5–7 ms

Cameras—

Modelyolov7-320

Version0.15.2

AcceleratorNvidia RTX 3060price not tracked
Host—
CPU—

Aug 2025GitHub #19741

Nvidia RTX 3060

ONNX (Nvidia / GPU)

Speed10–14 ms

Cameras—

Modelyolov9-t

Version0.16

AcceleratorNvidia RTX 3060price not tracked
Host—
CPU—

Aug 2025GitHub #19741

Nvidia RTX 3060

TensorRT (Nvidia, legacy) · avg 10 ms (40 ms with onnx)

Speed10 ms

Cameras—

Modelyolov7 trt

Version0.16.0-beta3-tensorrt

AcceleratorNvidia RTX 3060price not tracked
Host—
CPUAMD Ryzen 5 3600

Aug 2025GitHub #19741

Intel Arc A380

OpenVINO (Intel)

Frigate docs

Speed27 ms

Cameras—

Modelrf-detr 448

Version—

AcceleratorIntel Arc A380$159.99
Host—
CPU—

checked 2026-10-19Frigate docs

Intel Arc A380

OpenVINO (Intel)

Frigate docs

Speed20 ms

Cameras—

Modelrf-detr 336

Version—

AcceleratorIntel Arc A380$159.99
Host—
CPU—

checked 2026-10-19Frigate docs

3 other builds use the same device; the rest share the rf-detr model family.