Best home rig for Llama 3 70B
Older model. For current picks see 2025–2026 models.
Real, measured generation speeds for Llama 3 70B on home hardware, fastest first. Different quants/engines aren't directly comparable.
| Build | Quant | Engine | tok/s | Cost | |
|---|---|---|---|---|---|
| Dual RTX 3090 on Ryzen 3700X (X470) | AWQ 4-bit | vLLM (TP=2) | 21.2 | source ⚠ unchecked | |
| Dual RTX 3090 on Ryzen 3700X (X470) | Q4_K_M | Ollama | 16.95 | source ⚠ unchecked |
SPONSOR
Your brand here. One sponsor slot on this page, shown to people choosing AI hardware. Sponsor RigBase →
Your brand here. One sponsor slot on this page, shown to people choosing AI hardware. Sponsor RigBase →
Rent before you buy: try this model on a cloud GPU for a few dollars first: RunPod
FAQ
What is the fastest measured hardware for Llama 3 70B?
2x RTX 3090: 21.2 tok/s generation (vLLM (TP=2), AWQ 4-bit), from https://www.reddit.com/r/LocalLLaMA/comments/1d5vxe7/so_i_bought_second_3090_here_are_my_results_llama/.
Related: what GPU do you need for 2026 models · all builds