Qwen3.5-0.8B on Jetson Orin NX 16GB — Q8_0, llama.cpp vision-language inference benchmark

Open this configuration in the VLM Inference Benchmark Explorer

Model	Qwen3.5-0.8B
Parameters	0.9B
Device	Jetson Orin NX 16GB
Quantization	Q8_0
Image size	720p
Images per camera	1
Max Cameras (response ≤ 3 s)	2
Engine	llama.cpp

Cameras	Response (s)	TPS (tok/s)	Target
1	0.97	43.2	≤3 s
2	1.50	44.3	≤3 s
4	5.91	50.0	>3 s
8	12.9	50.1	>3 s

Images per request	Response (s)	TPS (tok/s)	Target
1	0.97	43.2	≤3 s
2	1.90	42.7	≤3 s
4	3.72	41.8	>3 s
8	7.43	39.7	>3 s
16	15.2	36.7	>3 s
32	31.8	31.8	>3 s

Qwen3.5-0.8B (0.9B parameters), served in Q8_0 format with llama.cpp on Jetson Orin NX 16GB, starts answering a single camera that sends one 720p image after 0.97 s, then writes 43.2 tok/s.

With 32 images per camera at 720p, the answer starts after 31.75 s. With one 2K image, after 6.5 s.

At the default target of an answer starting within 3 s (720p, one image per camera), it keeps up with 2 cameras at once.