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

camembert-base

by almanach

111M params · fill-mask · 99 likes · 942.7k downloads

camembert-base is a 111M parameter model. At Q4 quantization it requires 0GB of VRAM. It runs comfortably on GeForce RTX 4090 (11802 tok/s), GeForce RTX 5090 (17699 tok/s), M4 Max 128GB (4319 tok/s).

Inference providers

Provider$/1M in$/1M outThroughput
HF Inference

GPU compatibility

GPUVRAMQ4 DecodeVerdict
GeForce RTX 409024GB11802 tok/scomfortable
GeForce RTX 509032GB17699 tok/scomfortable
M4 Max 128GB128GB4319 tok/scomfortable
M4 Pro 48GB48GB2159 tok/scomfortable
M4 Pro 24GB24GB2159 tok/scomfortable
A100 PCIe 80 GB80GB21624 tok/scomfortable
H100 SXM5 80 GB80GB43492 tok/scomfortable
GeForce RTX 309024GB10435 tok/scomfortable
Radeon RX 7900 XTX24GB8629 tok/scomfortable
GeForce RTX 408016GB8376 tok/scomfortable
Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0