Llama 3 8B

8B · Family: Llama 3

Meta's 8B-parameter base (pretrained) Llama 3 language model, released in 2024. Not instruction-tuned.

✓ Active✓ Public access⚖ Open weightsLLM📁 Llama 3

Context window

tokens

Parameters

parameters

Release date

18 April 2024

🏢Meta AIProducer

Access:DownloadAPIDeployment:💻 Local☁ Cloud

Overview

Llama 3 8B (Meta-Llama-3-8B) is Meta's 8B-parameter base, pretrained language model from the Llama 3 family, released on April 18, 2024. It is the base variant rather than the instruction-tuned (Instruct) version — it has not undergone instruction tuning or RLHF.

The model uses an optimized auto-regressive decoder-only transformer architecture with grouped-query attention (GQA) and a 128K-token tokenizer vocabulary. It was pretrained on over 15T tokens of publicly available data, with a knowledge cutoff of March 2023. The context length is 8,192 tokens (8K).

The model takes text input and generates text and code. Meta lists commercial and research use in English. Weights are available for download under the Meta Llama 3 Community License.

Classification

LLM

Family: Llama 3

Access & deployment

DownloadAPI

LocalCloud

Weights: Open weights

Key parameters

📏 Context: 8K

🧩 Parameters: 8B

✓ Fine-tuning

📥 Input: text

Technical specification

Context window

tokens

Parameters

parameters

Knowledge cutoff

1 Mar 2023

Knowledge boundary

License

Meta Llama 3 Community License

Hardware requirements

The 8B model fits on a single consumer/server GPU; in BF16 it needs roughly 16 GB of memory, less when quantized.

Features:✓ Fine-tuning

Modalities

⬇ Input

text

⬆ Output

textcode

Capabilities and applications

Native model capabilities

Coding

Generating, analysing and modifying code in many programming languages. Covers writing functions, debugging, refactoring, code review, and creating tests. Measured by benchmarks such as HumanEval and SWE-bench.

Category: coding