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

OLMo-7B

by allenai

6.9B params · text-generation · 652 likes · 2.3k downloads

OLMo-7B is a 6.9B parameter model. At Q4 quantization it requires 3GB of VRAM. It runs comfortably on GeForce RTX 4090 (190 tok/s), GeForce RTX 5090 (285 tok/s), M4 Max 128GB (69 tok/s).

GPU compatibility

GPUVRAMQ4 DecodeVerdict
GeForce RTX 409024GB190 tok/scomfortable
GeForce RTX 509032GB285 tok/scomfortable
M4 Max 128GB128GB69 tok/scomfortable
M4 Pro 48GB48GB34 tok/scomfortable
M4 Pro 24GB24GB34 tok/scomfortable
A100 PCIe 80 GB80GB349 tok/scomfortable
H100 SXM5 80 GB80GB702 tok/scomfortable
GeForce RTX 309024GB168 tok/scomfortable
Radeon RX 7900 XTX24GB139 tok/scomfortable
GeForce RTX 408016GB135 tok/scomfortable
Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0