Llama-3.2-3B

Name: Llama-3.2-3B
Author: Facebook AI研究实验室

Foundation modelLlama-3.2

Llama-3.2-3B

Release date: 2024-09-18Updated: 2024-12-08 21:36:26894

Live demoGitHub Hugging Face Compare

Parameters

3.2B

Context length

128K

Chinese support

Not supported

Reasoning ability

Llama-3.2-3B is an AI model published by Facebook AI研究实验室, released on 2024-09-18, for Foundation model, with 3.2B parameters, and 128K context length, requiring about 6.43GB storage, under the LLAMA 3.1 COMMUNITY LICENSE AGREEMENT license, with a 54.75 score on MMLU.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Llama-3.2-3B

Model basics

Reasoning traces

Not supported

Thinking modes

Thinking modes not supported

Context length

128K tokens

Max output length

No data

Model type

Foundation model

Modality (in / out)

No data

Release date

2024-09-18

Model file size

6.43GB

MoE architecture

Total params / Active params

3.2B / N/A

Knowledge cutoff

No data

Llama-3.2-3B

Open source & experience

Code license

LLAMA 3.1 COMMUNITY LICENSE AGREEMENT

Weights license

LLAMA 3.1 COMMUNITY LICENSE AGREEMENT- 免费商用授权

GitHub repo

https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/MODEL_CARD.md

Hugging Face

https://huggingface.co/meta-llama/Llama-3.2-3B

Live demo

No live demo

Llama-3.2-3B

Official resources

Paper

Llama 3.2-Model Card

DataLearnerAI blog

No blog post yet

Llama-3.2-3B

API details

API speed

No data

No public API pricing yet.

Llama-3.2-3B

Benchmark Results

Llama-3.2-3B currently shows benchmark results led by GSM8K (23 / 26, score 34), BBH (18 / 20, score 46.80), GPQA Diamond (172 / 179, score 26.60). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.