LLM EXPLORER 61,405 MODELS INDEXED

Llama 3.2 3B Instruct by unsloth

By unsloth · 186207 downloads
HF Repo ↗ Copy URL

Llama 3.2 3B Instruct is an open-source language model by unsloth. Features: 3b LLM, VRAM: 6.5GB, Context: 128K, License: llama3.2, Instruction-Based, LLM Explorer Score: 0.34, ELO: 1167.

Base model:finetune:meta-llama... Base model:meta-llama/llama-3....   Conversational   Deploy:azure   En   Endpoints compatible   Facebook   Instruct   Llama   Llama-3   Meta   Region:us   Safetensors   Sharded   Tensorflow   Unsloth
Model Card on HF 🤗: https://huggingface.co/unsloth/Llama-3.2-3B-Instruct 

Llama 3.2 3B Instruct Benchmarks

LMSys ELO 1167 vs 1272 (so35) -8.3%
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

Llama 3.2 3B Instruct Parameters and Internals

Model Type 
auto-regressive, transformer, text generation
Use Cases 
Areas:
multilingual dialogue use cases
Primary Use Cases:
agentic retrieval, summarization tasks
Additional Notes 
Future versions may be released that improve model capabilities and safety.
Supported Languages 
English (official), German (official), French (official), Italian (official), Portuguese (official), Hindi (official), Spanish (official), Thai (official)
Training Details 
Methodology:
supervised fine-tuning, reinforcement learning with human feedback (RLHF)
Model Architecture:
optimized transformer architecture
Input Output 
Input Format:
text
Accepted Modalities:
text
Output Format:
text
Release Notes 
Date:
Sept 25, 2024
Notes:
This is a static model trained on an offline dataset.
LLM NameLlama 3.2 3B Instruct
Repository 🤗https://huggingface.co/unsloth/Llama-3.2-3B-Instruct 
Base Model(s)  meta-llama/Llama-3.2-3B-Instruct   meta-llama/Llama-3.2-3B-Instruct
Model Size3b
Required VRAM6.5 GB
Updated2026-08-08
Maintainerunsloth
Model Typellama
Instruction-BasedYes
Model Files  5.0 GB: 1-of-2   1.5 GB: 2-of-2
Supported Languagesen
Model ArchitectureLlamaForCausalLM
Licensellama3.2
Context Length131072
Model Max Length131072
Transformers Version4.52.4
Tokenizer ClassPreTrainedTokenizer
Padding Token<|finetune_right_pad_id|>
Vocabulary Size128256
Torch Data Typebfloat16

Quantized Models of the Llama 3.2 3B Instruct

Model
Likes
Downloads
VRAM
ReWiz Llama 3.2 3B33416 GB

Best Alternatives to Llama 3.2 3B Instruct

Best Alternatives
Context / RAM
Downloads
Likes
Llama 3.2 3B Instruct128K / 6.5 GB18295722368
...ini Math345 GroupB Llama32 End128K / 6.5 GB7550
...ama 3.2 3B TechWriter Instruct128K / 6.4 GB2260
HyLo Llama 6MLA22GDN 8K SFT128K / 17.1 GB421
Llama 3.2 3B Instruct128K / 6.5 GB9919
Llama3.2 3B Manumit V1128K / 6.4 GB1350
Alloma 3B Instruct128K / 6.5 GB554566
... V7 SFT 10voicebot Loopcleaned128K / 6.5 GB2290
Llama 3.2 3B RP Toxic Fuse128K / 6.4 GB142
DeepSeek R1 Distill Llama 3B128K / 6.5 GB24716
Note: green Score (e.g. "73.2") means that the model is better than unsloth/Llama-3.2-3B-Instruct.