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

t5-large

by google-t5

738M params · translation · 250 likes · 343.9k downloads

t5-large is a 738M parameter model. At Q4 quantization it requires 0GB of VRAM. It runs comfortably on GeForce RTX 4090 (1779 tok/s), GeForce RTX 5090 (2669 tok/s), M4 Max 128GB (651 tok/s).

Inference providers

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

GPU compatibility

GPUVRAMQ4 DecodeVerdict
GeForce RTX 409024GB1779 tok/scomfortable
GeForce RTX 509032GB2669 tok/scomfortable
M4 Max 128GB128GB651 tok/scomfortable
M4 Pro 48GB48GB325 tok/scomfortable
M4 Pro 24GB24GB325 tok/scomfortable
A100 PCIe 80 GB80GB3261 tok/scomfortable
H100 SXM5 80 GB80GB6559 tok/scomfortable
GeForce RTX 309024GB1573 tok/scomfortable
Radeon RX 7900 XTX24GB1301 tok/scomfortable
GeForce RTX 408016GB1263 tok/scomfortable
Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0