What are the hardware requirements for Llama 3 Smaug 8B AWQ?

Llama 3 Smaug 8B AWQ requires approximately 5.8 GB of VRAM and supports a context window of 8K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Llama 3 Smaug 8B AWQ and how large is it?

Llama 3 Smaug 8B AWQ is developed by solidrust, a model with 8b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Llama 3 Smaug 8B AWQ perform on standard benchmarks?

Llama 3 Smaug 8B AWQ has the following published scores: MMLU 65.77. Compare against reference models on this page or on the LLM Explorer leaderboards.

Llama 3 Smaug 8B AWQ by solidrust — VRAM 5.8GB, 8K context

Name: Llama 3 Smaug 8B AWQ
Author: solidrust

Llama 3 Smaug 8B AWQ is an open-source language model by solidrust. Features: 8b LLM, VRAM: 5.8GB, Context: 8K, License: llama2, Quantized, LLM Explorer Score: 0.14, Arc: 59.6, HellaSwag: 80, MMLU: 65.8, GSM8K: 55.9.

4-bit Autotrain compatible Awq Base model:abacusai/llama-3-sm... Base model:quantized:abacusai/... Conversational Dataset:anon8231489123/sharegp... Dataset:aqua rat Dataset:m-a-p/codefeedback-fil... Dataset:microsoft/orca-math-wo... Endpoints compatible Llama Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/solidrust/Llama-3-Smaug-8B-AWQ

Llama 3 Smaug 8B AWQ Parameters and Internals

Model Type

text-generation

Additional Notes

This model was built using the Smaug recipe for improving performance on real-world multi-turn conversations. AWQ is an efficient, accurate, and blazing-fast low-bit weight quantization method, supporting 4-bit quantization. AWQ models are currently supported on Linux and Windows with NVidia GPUs only. macOS users should use GGUF models instead.

LLM Name	Llama 3 Smaug 8B AWQ
Repository 🤗	https://huggingface.co/solidrust/Llama-3-Smaug-8B-AWQ
Base Model(s)	Llama 3 Smaug 8B abacusai/Llama-3-Smaug-8B
Model Size	8b
Required VRAM	5.8 GB
Updated	2026-08-08
Maintainer	solidrust
Model Type	llama
Model Files	4.7 GB: 1-of-2 1.1 GB: 2-of-2
AWQ Quantization	Yes
Quantization Type	awq
Model Architecture	LlamaForCausalLM
License	llama2
Context Length	8192
Model Max Length	8192
Transformers Version	4.38.2
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<\|end_of_text\|>
Vocabulary Size	128256
Torch Data Type	float16

Rank the Llama 3 Smaug 8B AWQ capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Llama 3 Smaug 8B AWQ

Best Alternatives	Context / RAM	Downloads	Likes
...8B Instruct Gradient 1048K AWQ	1024K / 5.8 GB	8	0
...radient 1048K AWQ 4bit Smashed	1024K / 5.8 GB	7	1
...Instruct 262K AWQ 4bit Smashed	256K / 5.8 GB	5	4
...ta Llama 3 8B Instruct 64K AWQ	64K / 5.8 GB	6	0
... Instruct 8B 32K V0.1 4bit AWQ	64K / 5.8 GB	13	0
Llama 3 8B Instruct AWQ	8K / 5.8 GB	25458	31
Meta Llama 3 8B Instruct AWQ	8K / 5.8 GB	6	0
Meta Llama 3 8B Instruct AWQ	8K / 5.8 GB	563	0
Meta Llama 3 8B Instruct AWQ	8K / 5.8 GB	7	0
Meta Llama 3 8B Instruct AWQ	8K / 5.8 GB	401	5

Note: green Score (e.g. "73.2") means that the model is better than solidrust/Llama-3-Smaug-8B-AWQ.