can-i-run-this-llm

GPU comparison

Mac Studio M3 Ultra (96GB) vs NVIDIA RTX 4070 (12GB) for local LLMs

How the two stack up for running open-source LLMs locally — memory, bandwidth, price, and how many of the 34 tracked models each one can run.

Mac Studio M3 Ultra (96GB)NVIDIA RTX 4070 (12GB)
Memory96.0 GB12.0 GB
Bandwidth819 GB/s504 GB/s
Price (approx)$5,299$549
LLMs it runs30 of 3413 of 34
Best model it runsGemma 4 31B · 15–22 tok/sDevstral Small 2 24B · 35–53 tok/s

The Mac Studio M3 Ultra (96GB) runs 17 more of the tracked models (30 vs 13), thanks to its 96.0 GB of memory. The Mac Studio M3 Ultra (96GB) has more memory bandwidth (819 GB/s), so it generates tokens faster at the same model and quant.

What the Mac Studio M3 Ultra (96GB) runs →What the NVIDIA RTX 4070 (12GB) runs →Open the calculator →