Models /Llama 3.1 8B Instruct /Can I run it?

Compatibility check

Can I run Llama 3.1 8B Instruct on GeForce RTX 4070 SUPER?

Yes — it runs.best quant:

Decode speed

83 tok/sest

Usable context

66K

Memory

12 GB

Fits in memory83 tok/sest · 66K ctx

Fits in memory71 tok/sest · 66K ctx

Fits in memoryest47 tok/sest · 8K ctx

What is quantisation? →

Llama 3.1 8B Instruct

8B · full specs & benchmarks →

GeForce RTX 4070 SUPER

12 GB · everything it runs →