# Llama 2 7B Chat: hardware requirements
> Llama 2 7B Chat (Meta) is a 6.7B dense model with a 4K-token context window, released under the Llama 2 Community License. It needs about 6.9GB of memory at Q4_K_M with 8K context.
Llama 2 7B was Meta's July 2023 entry-level open model and the default local model of its era. Today it is useful mainly as a historical baseline: current 4B to 9B models beat it by a wide margin on reasoning, coding, and instruction following. It runs on almost anything, needing about 4GB at Q4_K_M plus up to 2GB of KV cache at its 4K context limit, so any 8GB GPU or 8GB Apple silicon Mac handles it. For new projects, pick Qwen3.5 9B or Gemma 4 E4B instead.

## Recommended hardware
- **Minimum GPU:** RTX 5060 (8GB), Q4_K_M, ~81 tokens/sec
- **Best value:** RTX 5060 (8GB), Q4_K_M, ~81 tokens/sec
- **Fastest single GPU:** RTX 5090 (32GB), Q8_0, ~165 tokens/sec
- **Mac / unified memory:** Ryzen AI Max+ 395 (128GB), Q8_0, ~26 tokens/sec

## Memory by quantization
| Quant | Weights | Total @ 8K ctx | Total @ 32K ctx |
| --- | --- | --- | --- |
| Q2_K | 2.5GB | 5.4GB | - |
| Q3_K_M | 3.3GB | 6.2GB | - |
| Q4_K_M | 4.1GB | 6.9GB | - |
| Q5_K_M | 4.8GB | 7.7GB | - |
| Q6_K | 5.5GB | 8.4GB | - |
| Q8_0 | 7.2GB | 10.1GB | - |
| BF16 | 13.5GB | 16.6GB | - |

## GPUs that run it fully in memory (fastest first)
| GPU | Memory | Quant | Speed | Price |
| --- | --- | --- | --- | --- |
| RTX 5090 | 32GB | Q8_0 | ~165 tok/s | ~$4,300 |
| RTX PRO 6000 | 96GB | Q8_0 | ~165 tok/s | ~$16,000 |
| RTX PRO 5000 72GB | 72GB | Q8_0 | ~130 tok/s | ~$9,100 |
| RTX 4090 | 24GB | Q8_0 | ~100 tok/s | ~$2,500 used |
| RTX 5080 | 16GB | Q8_0 | ~97 tok/s | ~$1,600 |
| RTX 6000 Ada | 48GB | Q8_0 | ~97 tok/s | ~$4,750 used |
| RTX 5070 Ti | 16GB | Q8_0 | ~91 tok/s | ~$1,200 |
| RTX 3090 | 24GB | Q8_0 | ~86 tok/s | ~$1,050 used |
| M5 Ultra 512GB | 512GB | Q8_0 | ~83 tok/s | n/a |
| M5 Ultra 256GB | 256GB | Q8_0 | ~83 tok/s | ~$9,499 |
| RTX 5060 | 8GB | Q4_K_M | ~81 tok/s | ~$450 |
| RX 7900 XTX | 24GB | Q8_0 | ~77 tok/s | ~$1,050 used |
| RTX 4080 Super | 16GB | Q8_0 | ~76 tok/s | ~$1,150 used |
| V100 32GB | 32GB | Q8_0 | ~74 tok/s | ~$700 used |
| RTX A6000 | 48GB | Q8_0 | ~72 tok/s | ~$4,400 used |
| RTX 4070 Ti Super | 16GB | Q8_0 | ~70 tok/s | ~$900 used |
| RTX 5070 | 12GB | Q8_0 | ~70 tok/s | ~$800 |
| M3 Ultra 512GB | 512GB | Q8_0 | ~57 tok/s | $9,499 MSRP |
| MI50 32GB | 32GB | Q8_0 | ~57 tok/s | ~$600 used |
| R9700 | 32GB | Q8_0 | ~53 tok/s | ~$1,800 |
| RX 9070 XT | 16GB | Q8_0 | ~53 tok/s | ~$800 |
| RTX 4070 | 12GB | Q8_0 | ~53 tok/s | ~$560 used |
| M5 Max 128GB | 128GB | Q8_0 | ~51 tok/s | ~$5,100 |
| RTX 5060 Ti 16GB | 16GB | Q8_0 | ~48 tok/s | ~$620 |
| M4 Max 128GB | 128GB | Q8_0 | ~46 tok/s | $3,499 MSRP |

## How to run it
- Ollama: `ollama run llama2:7b`
- llama.cpp: `llama-server -hf TheBloke/Llama-2-7B-Chat-GGUF:Q4_K_M -c 8192 -ngl 99`
- Weights: https://huggingface.co/meta-llama/Llama-2-7b-chat-hf

## Strengths
- Tiny footprint: about 4GB at Q4_K_M, fast even on old 6GB and 8GB GPUs
- Supported by every runtime, fine-tuning tool, and tutorial
- Large archive of community fine-tunes for niche tasks

## Weaknesses
- 4K context window is too short for documents, codebases, or long chats
- Far behind 2025 and 2026 small models on reasoning, coding, and factual accuracy
- No grouped-query attention, so its KV cache per token is four times larger than Llama 3.1 8B's
- Llama 2 license adds use restrictions and a 700M monthly user cap
Full page: https://llamabuilds.ai/models/llama-2-7b · Hardware guide: https://llamabuilds.ai/models/llama-2-7b/requirements

---
Source: https://llamabuilds.ai/models/llama-2-7b · Updated October 14, 2026 · Llama Builds. Speeds marked ~ are estimates from our bandwidth model (https://llamabuilds.ai/methodology); measured results are labeled as such.