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

starcoder2-15b

by bigcode

31.9B params · text-generation · 661 likes · 22.5k downloads

starcoder2-15b is a 31.9B parameter model. At Q4 quantization it requires 16GB of VRAM. It runs comfortably on GeForce RTX 4090 (41 tok/s), GeForce RTX 5090 (61 tok/s), A100 PCIe 80 GB (75 tok/s).

GPU compatibility

GPUVRAMQ4 DecodeVerdict
GeForce RTX 409024GB41 tok/scomfortable
GeForce RTX 509032GB61 tok/scomfortable
M4 Max 128GB128GB15 tok/stight
M4 Pro 48GB48GB7 tok/stight
M4 Pro 24GB24GB7 tok/stight
A100 PCIe 80 GB80GB75 tok/scomfortable
H100 SXM5 80 GB80GB151 tok/scomfortable
GeForce RTX 309024GB36 tok/scomfortable
Radeon RX 7900 XTX24GB30 tok/scomfortable
M4 Max 64GB64GB15 tok/stight
Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0