DetectFPS
Submit a build

Nvidia RTX 5070 Ti · yolov9-s 320 · 20 cameras

Frigate 0.17.2Community build
Posted Aug 13, 2026·GitHub #23978 ·db-155
I run this too, add my numbers Opens the submit form prefilled with this hardware, model and version
Inference
10.5 ms
as reported
Est. capacity
~95 det/s
1000 ÷ 10.5 ms · estimate
Cameras
20
About 20-25 cameras (20 recorded).
CPU
—
not reported
Power
—
not reported

Parts

prices show only once verified by hand
Accelerator
Nvidia RTX 5070 Ti
ONNX (Nvidia / GPU) · 1 build on DetectFPS
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
Intel Core i7-14700KF
as reported
—
Cameras
20 cameras
resolution not reported
—
Tracked totalno verified prices yet

Setup

Detector
ONNX (Nvidia / GPU)
Detector instances
—
Model, as reported
yolov9-s 320
Model family · input
yolov9-s · 320
Camera resolution
—
Cameras
20
Frigate version
0.17.2 (0.17 series, 1 behind newest)
OS
—

Numbers

dashes mean the source didn't say
Inference
10.5 ms
Inference range
—
Total fps
—
CPU
—
Power
—
Used in stats
Yes (the reading is valid despite the flag)

Notes and flags

Known issue. ONNX detector got stuck with CUDA Graphs on; fixed by disabling CUDA Graphs.

About 20-25 cameras (20 recorded).

Source

Where
Date, as the source wrote it
2026-08-13 (day precision)
Added to DetectFPS
from a public post; numbers copied as written, not re-tested

Similar builds

Same device, then the same model family

Intel UHD 730 (12th-14th gen iGPU)
OpenVINO (Intel) · Detector CPU 50-75%.
Speed15–20 ms
Cameras1
Modelyolov9-s 320
Version—
  • AcceleratorIntel UHD 730 (12th-14th gen iGPU)price not tracked
  • Host—
  • CPUIntel Core i5-14400
Nov 2025GitHub #20991
Intel Arc A310
OpenVINO (Intel) · ~20 ms; no Resizable BAR support
Speed20 ms
Cameras15
Modelyolov9-s 320
Version—
  • AcceleratorIntel Arc A310price not tracked
  • HostHPE ProLiant DL380 Gen10price not tracked
  • CPU—
Mar 1, 2026GitHub #22189
Speed30 ms
Cameras3
Modelyolov9-s 320
Version—
  • AcceleratorIntel HD 530 (6th-gen iGPU)price not tracked
  • HostCirca-2015 PC (6th-gen Intel)price not tracked
  • CPU—
Oct 8, 2025fuzznotes.com
Intel NPU (Core Ultra)
OpenVINO (Intel)
Frigate docs
Speed11 ms
Cameras—
Modelyolov9-s 320
Version—
  • AcceleratorIntel NPU (Core Ultra)price not tracked
  • Host—
  • CPU—
checked 2026-10-19Frigate docs
Nvidia RTX 3050
ONNX (Nvidia / GPU)
Frigate docs
Speed10 ms
Cameras—
Modelyolov9-s 320
Version—
  • AcceleratorNvidia RTX 3050price not tracked
  • Host—
  • CPU—
checked 2026-10-19Frigate docs
Apple M4
Apple Silicon
Frigate docs
Speed10 ms
Cameras—
Modelyolov9-s 320
Version—
  • SoCApple M4price not tracked
  • HostApple M4price not tracked
  • CPU—
checked 2026-10-19Frigate docs

No other build uses this exact device yet; these share the yolov9-s model family (34 builds), nearest speeds first.