can-i-run-this-llm

GPU comparison

NVIDIA RTX 4070 (12GB) vs NVIDIA RTX 4090 (24GB) for local LLMs

How the two stack up for running open-source LLMs locally — memory, bandwidth, price, and how many of the 18 tracked models each one can run.

NVIDIA RTX 4070 (12GB)NVIDIA RTX 4090 (24GB)
Memory12.0 GB24.0 GB
Bandwidth504 GB/s1008 GB/s
Price (approx)$549$1,599
LLMs it runs8 of 1815 of 18
Best model it runsGemma 4 12B · 42–63 tok/sGemma 4 31B · 33–50 tok/s

The NVIDIA RTX 4090 (24GB) runs 7 more of the tracked models (15 vs 8), thanks to its 24.0 GB of memory. The NVIDIA RTX 4090 (24GB) has more memory bandwidth (1008 GB/s), so it generates tokens faster at the same model and quant.

What the NVIDIA RTX 4070 (12GB) runs →What the NVIDIA RTX 4090 (24GB) runs →Open the calculator →