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

Qwen3.5-2B

by Qwen

2.3B params · image-text-to-text · 314 likes · 1.6M downloads

Qwen3.5-2B is a 2.3B parameter model. At Q4 quantization it requires 1GB of VRAM. It runs comfortably on GeForce RTX 4090 (577 tok/s), GeForce RTX 5090 (865 tok/s), M4 Max 128GB (211 tok/s).

Inference providers

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

GPU compatibility

GPUVRAMQ4 DecodeVerdict
GeForce RTX 409024GB577 tok/scomfortable
GeForce RTX 509032GB865 tok/scomfortable
M4 Max 128GB128GB211 tok/scomfortable
M4 Pro 48GB48GB105 tok/scomfortable
M4 Pro 24GB24GB105 tok/scomfortable
A100 PCIe 80 GB80GB1057 tok/scomfortable
H100 SXM5 80 GB80GB2127 tok/scomfortable
GeForce RTX 309024GB510 tok/scomfortable
Radeon RX 7900 XTX24GB422 tok/scomfortable
GeForce RTX 408016GB409 tok/scomfortable
Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0