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

gemma-7b-it

by google

8.5B params · text-generation · 1.2k likes · 19.9k downloads

gemma-7b-it is a 8.5B parameter model. At Q4 quantization it requires 4GB of VRAM. It runs comfortably on GeForce RTX 4090 (153 tok/s), GeForce RTX 5090 (230 tok/s), M4 Max 128GB (56 tok/s).

Inference providers

Provider$/1M in$/1M outThroughput
Featherless

GPU compatibility

GPUVRAMQ4 DecodeVerdict
GeForce RTX 409024GB153 tok/scomfortable
GeForce RTX 509032GB230 tok/scomfortable
M4 Max 128GB128GB56 tok/scomfortable
M4 Pro 48GB48GB28 tok/stight
M4 Pro 24GB24GB28 tok/stight
A100 PCIe 80 GB80GB281 tok/scomfortable
H100 SXM5 80 GB80GB566 tok/scomfortable
GeForce RTX 309024GB135 tok/scomfortable
Radeon RX 7900 XTX24GB112 tok/scomfortable
GeForce RTX 408016GB109 tok/scomfortable
Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0