vram.run Models Hardware Providers Cloud State of Inference
API provider data is live · Hardware & cloud pricing curated 2026-02-23

gemma-3n-E4B-it

by google

7.8B params · image-text-to-text · 918 likes · 21.0k downloads

gemma-3n-E4B-it is a 7.8B parameter model. At Q4 quantization it requires 4GB of VRAM. It runs comfortably on GeForce RTX 4090 (167 tok/s), GeForce RTX 5090 (250 tok/s), M4 Max 128GB (61 tok/s).

Inference providers

Provider$/1M in$/1M outThroughput
Together AI54 tok/s

GPU compatibility

GPUVRAMQ4 DecodeVerdict
GeForce RTX 409024GB167 tok/scomfortable
GeForce RTX 509032GB250 tok/scomfortable
M4 Max 128GB128GB61 tok/scomfortable
M4 Pro 48GB48GB30 tok/scomfortable
M4 Pro 24GB24GB30 tok/scomfortable
A100 PCIe 80 GB80GB306 tok/scomfortable
H100 SXM5 80 GB80GB616 tok/scomfortable
GeForce RTX 309024GB147 tok/scomfortable
Radeon RX 7900 XTX24GB122 tok/scomfortable
GeForce RTX 408016GB118 tok/scomfortable
Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0