Meta
Llama 3.1 405B (405B parameters) requires approximately 256.2 GB of VRAM with Q4_K_M quantization. For the best balance of quality and speed, we recommend hardware with at least 295 GB of VRAM.
Get started
— copy & paste to run locallyCopy-paste commands to run Llama 3.1 405B on your machine.
Run
ollama run llama3.1:405bQuick specs
About this model
Related models
Quick picks
Best hardware
Run this model
Quantization options
No hardware detected — fit column shows raw VRAM estimates
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 158.0 GB | Low | — |
Q3_K_S | 3 | 198.5 GB | Low | — |
NVFP4 | 4 | 226.8 GB | Medium | — |
Q4_K_M | 4 | 247.1 GB | Medium | — |
Q5_K_M | 5 | 291.6 GB | High | — |
Q6_K | 6 | 332.1 GB | High | — |
Q8_0 | 8 | 433.4 GB | Very High | — |
F16 | 16 | 830.2 GB | Maximum | — |
Quality benchmarks
Coding
Reasoning
General
Source: official · 2024-07-23
Hardware compatibility
Computing compatibility...
Memory breakdown
Frequently asked questions
Llama 3.1 405B (405B parameters) requires approximately 256.2 GB of VRAM with Q4_K_M quantization. Lower quantizations like Q4_K_M use less memory but may reduce quality.
The recommended quantization for Llama 3.1 405B is Q4_K_M, which offers the best balance between model quality and memory efficiency. Higher quantizations preserve more quality but require more VRAM.
The top recommended hardware for Llama 3.1 405B: AMD Instinct MI350X 288GB (score: 84/100), AMD Instinct MI325X 256GB (score: 71/100). These provide the best combination of memory, bandwidth, and compute for running this model locally.
Yes, Llama 3.1 405B is well-suited for chat as well as reasoning, coding, agentic. It was designed with these use cases in mind.
See also