LensVLM-9B on Jetson AGX Orin 32GB Developer Kit — Q8_0, llama.cpp vision-language inference benchmark

Open this configuration in the VLM Inference Benchmark Explorer

Model	LensVLM-9B
Parameters	9.4B
Device	Jetson AGX Orin 32GB Developer Kit
Quantization	Q8_0
Image size	720p
Images per camera	1
Max Cameras (response ≤ 3 s)	1
Engine	llama.cpp

Cameras	Response (s)	TPS (tok/s)	Target
1	2.73	17.5	≤3 s
2	3.17	9.8	>3 s
4	4.56	7.8	>3 s
8	21.4	8.5	>3 s

Images per request	Response (s)	TPS (tok/s)	Target
1	2.73	17.5	≤3 s
2	5.27	17.4	>3 s
4	10.3	17.2	>3 s
8	20.5	16.9	>3 s
16	41.1	16.3	>3 s
32	83.4	15.1	>3 s

LensVLM-9B (9.4B parameters), served in Q8_0 format with llama.cpp on Jetson AGX Orin 32GB Developer Kit, starts answering a single camera that sends one 720p image after 2.73 s, then writes 17.5 tok/s.

With 32 images per camera at 720p, the answer starts after 83.45 s. With one 2K image, after 20.21 s.

At the default target of an answer starting within 3 s (720p, one image per camera), it keeps up with one camera at a time.