Qwen3.5-9B on Jetson Orin NX 16GB — Q4_K_M, llama.cpp vision-language inference benchmark

Open this configuration in the VLM Inference Benchmark Explorer

Model	Qwen3.5-9B
Parameters	9.7B
Device	Jetson Orin NX 16GB
Quantization	Q4_K_M
Image size	720p
Images per camera	1
Max Cameras (response ≤ 3 s)	>3 s
Engine	llama.cpp

Cameras	Response (s)	TPS (tok/s)	Target
1	7.22	10.3	>3 s
2	11.3	8.1	>3 s
4	39.8	9.6	>3 s
8	83.6	9.6	>3 s

Images per request	Response (s)	TPS (tok/s)	Target
1	7.22	10.3	>3 s
2	14.3	10.2	>3 s
4	28.3	10.1	>3 s
8	56.6	9.8	>3 s
16	113.8	9.3	>3 s
32	231.0	8.4	>3 s

Qwen3.5-9B (9.7B parameters), served in Q4_K_M format with llama.cpp on Jetson Orin NX 16GB, starts answering a single camera that sends one 720p image after 7.22 s, then writes 10.3 tok/s.

With 32 images per camera at 720p, the answer starts after 231.05 s. With one 2K image, after 56.35 s.

At the default target of an answer starting within 3 s (720p, one image per camera), not even one camera is answered in time.