Compatibility check
Can I run Llama 3.1 8B Instruct on GeForce RTX 3060 8GB?
Yes — it runs.best quant: Q4_K_M
Decode speed
59 tok/sest
Usable context
16K
Memory
8 GB
Q4_K_MFits in memory59 tok/sest · 16K ctx
Q5_K_MSpills to system RAMest
Q8_0Spills to system RAMest