Meta
ChatCodingAI codingReasoningMulti-purpose70.6B
Chat

Llama 3.1 70B Instruct

Llama · Llama 3.1 Community License

Frontier-class open model. Approaches GPT-4 quality on many benchmarks. Requires significant VRAM; 48 GB+ recommended for usable quantizations. Well suited for serious local deployment.

Parameters
70.6B
Architecture
Dense
Context
131,072 tokens
Released
2024-07-23
Engines
llama.cpp, ollama, vLLM
Builder Tools
Cursor, Open WebUI

Parameters

70.6B

VRAM

47 GB

Context

128K

Formats

4

GPUs

22

Llama 3.1 70B Instruct (70.6B) requires 47 GB VRAM at recommended quality (Q5_K_M). On NVIDIA Grace Blackwell Ultra GB300, expect approximately 65 tok/s at Q5_K_M. For the best experience, High-End Home AI Server ($3,842) is recommended.

Source: OwnRig methodology

VRAM (Recommended)

47 GB

Quantization

Q5_K_M

File Size

42.4 GB

Max Context

128K tokens

Primary Use

Chat

Memory

VRAM Requirements

QualityQuantizationVRAMFile Size
recommendedQ5_K_M47 GB42.4 GB
efficientQ4_K_M39.5 GB35.3 GB
compressedQ3_K_M31.6 GB27.5 GB
compressedQ2_K24.5 GB21.2 GB
Scaling

Context Length Impact

KV cache VRAM at Q5_K_M quality. Longer context = more memory.

ContextKV CacheTotal VRAM
2K614 MB47.6 GBexceeds 24 GB
4K1.3 GB48.3 GBexceeds 24 GB
8K2.6 GB49.6 GBexceeds 24 GB
16K5.1 GB52.1 GBexceeds 24 GB
32K10.2 GB57.2 GBexceeds 24 GB
64K20.5 GB67.5 GBexceeds 24 GB
128K41 GB88 GBexceeds 24 GB

Compatible GPUs

22 devices
NVIDIA Grace Blackwell Ultra GB300Q5_K_M65 tok/sExcellent
Apple M4 Max (128GB Unified)Q5_K_M7 tok/sAcceptable
Apple M4 Max (64GB Unified)Q4_K_M8 tok/sAcceptable
Apple M4 Pro (48GB)Q4_K_M6 tok/sAcceptable
Apple M4 Ultra (192GB)Q5_K_M11 tok/sAcceptable
AMD Radeon Pro W7900Q4_K_M6 tok/sAcceptable
NVIDIA RTX PRO 6000 BlackwellQ5_K_M23 tok/sAcceptable
NVIDIA RTX PRO 6000 Blackwell Max-QQ5_K_M21 tok/sAcceptable
NVIDIA GeForce RTX 4090Q3_K_M5 tok/sMarginal
NVIDIA GeForce RTX 5090Q4_K_M9 tok/sMarginal
AMD Radeon RX 7900 XTXQ2_K1 tok/sMarginal
Apple M3 Pro (18GB Unified)Q2_KNot viable
NVIDIA GeForce RTX 3080 10GBQ2_KNot viable
NVIDIA GeForce RTX 4060 8GBQ2_KNot viable
NVIDIA RTX 4060 Laptop (40-60W)Q2_KNot viable
NVIDIA RTX 4070 Laptop (80-115W)Q2_KNot viable
NVIDIA GeForce RTX 4070 Ti 12GBQ2_KNot viable
AMD Radeon RX 7600Q2_KNot viable
AMD Radeon RX 9070Q2_KNot viable
AMD Radeon RX 9060 XT 16GBQ2_KNot viable
AMD Radeon RX 9060 XT 8GBQ2_KNot viable
NVIDIA GeForce RTX 5060 8GBQ2_KNot viable

Showing 22 of 22 entries

Builder Context

Llama 3.1 70B Instruct is commonly used with Cursor, Open WebUI. For an AI coding workflow, pair it with an embedding model like nomic-embed-text for local RAG.

Hardware

Recommended Builds

Complete PC builds that can run Llama 3.1 70B Instruct.

FAQ

Frequently Asked Questions

How much VRAM does Llama 3.1 70B Instruct need?
Llama 3.1 70B Instruct requires 47 GB VRAM at recommended quality (Q5_K_M). At lower quality settings, it can fit in as little as 24.5 GB.
What is the best GPU for Llama 3.1 70B Instruct?
The NVIDIA Grace Blackwell Ultra GB300 delivers the best performance for Llama 3.1 70B Instruct, achieving 65 tok/s at Q5_K_M with an excellent rating.
What quantization should I use for Llama 3.1 70B Instruct?
For the best quality, use Q5_K_M (47 GB VRAM). If your GPU has limited VRAM, Q2_K (24.5 GB) is the most efficient option with acceptable quality.
Is Llama 3.1 70B Instruct good for coding?
Yes. Llama 3.1 70B Instruct is used with Cursor, Open WebUI for local AI coding. For the best coding experience, pair it with an embedding model for local RAG.
All models

Data confidence: verified. Source

VRAM requirements are calculated from model parameters and may vary by inference engine, context length, and batch size. Performance estimates are based on community benchmarks and should be verified for your specific configuration.Llama is a trademark of its respective owner. OwnRig is not affiliated with or endorsed by the model creator.