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

SpatialLM-Llama-1B

by manycore-research

1.2B params · text-generation · 994 likes · 125 downloads

SpatialLM-Llama-1B is a 1.2B parameter model. At Q4 quantization it requires 1GB of VRAM. It runs comfortably on GeForce RTX 4090 (1052 tok/s), GeForce RTX 5090 (1578 tok/s), M4 Max 128GB (385 tok/s).

GPU compatibility

GPUVRAMQ4 DecodeVerdict
GeForce RTX 409024GB1052 tok/scomfortable
GeForce RTX 509032GB1578 tok/scomfortable
M4 Max 128GB128GB385 tok/scomfortable
M4 Pro 48GB48GB192 tok/scomfortable
M4 Pro 24GB24GB192 tok/scomfortable
A100 PCIe 80 GB80GB1928 tok/scomfortable
H100 SXM5 80 GB80GB3878 tok/scomfortable
GeForce RTX 309024GB930 tok/scomfortable
Radeon RX 7900 XTX24GB769 tok/scomfortable
GeForce RTX 408016GB747 tok/scomfortable
Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0