VLM Inference Benchmark Explorer

Gemma-4-E4B-it on Jetson AGX Orin 32GB — Q4_0, llama.cpp vision-language inference benchmark

Open this configuration in the VLM Inference Benchmark Explorer

ModelGemma-4-E4B-it
Parameters8.0B
DeviceJetson AGX Orin 32GB
QuantizationQ4_0
Image size720p
Images per camera1
Max Cameras (TTFT ≤ 3000 ms)2
Enginellama.cpp
CamerasTTFT (ms)TPS (tok/s)Target
1118326.2≤3000 ms
2178721.6≤3000 ms
4366118.7>3000 ms
8706512.4>3000 ms
Images per requestTTFT (ms)TPS (tok/s)Target
1118326.2≤3000 ms
3315025.8>3000 ms
5512725.7>3000 ms

Gemma-4-E4B-it (8.0B parameters), served in Q4_0 format with llama.cpp on Jetson AGX Orin 32GB, starts answering a single camera that sends one 720p image after 1,183 ms (time to first token, TTFT), then writes 26.2 tok/s.

With 5 images per camera at 720p, the answer starts after 5,127 ms. With one 2K image, after 3,432 ms.

At the default target of a mean TTFT within 3,000 ms (720p, one image per camera), it keeps up with 2 cameras at once.