What are the hardware requirements for Llama SciQ 4bits?

Llama SciQ 4bits 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 SciQ 4bits and how large is it?

Llama SciQ 4bits is developed by thewordsmiths, a model with 8.2b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Llama SciQ 4bits?

Llama SciQ 4bits is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Llama SciQ 4bits by thewordsmiths — VRAM 5.8GB, 8K context

Name: Llama SciQ 4bits
Author: thewordsmiths

Llama SciQ 4bits is an open-source language model by thewordsmiths. Features: 8.2b LLM, VRAM: 5.8GB, Context: 8K, License: apache-2.0, LLM Explorer Score: 0.13.

4-bit Bitsandbytes En Endpoints compatible Llama Region:us Safetensors Sharded Tensorflow Trl Unsloth

Model Card on HF 🤗: https://huggingface.co/thewordsmiths/Llama_SciQ_4bits

Llama SciQ 4bits Parameters and Internals

Model Type

text-generation-inference, transformers

Additional Notes

This model was trained 2x faster with Unsloth and Huggingface's TRL library.

Training Details

Methodology:

Used Unsloth and Huggingface's TRL library for faster training

LLM Name	Llama SciQ 4bits
Repository 🤗	https://huggingface.co/thewordsmiths/Llama_SciQ_4bits
Base Model(s)	thewordsmiths/llama_sciq thewordsmiths/llama_sciq
Model Size	8.2b
Required VRAM	5.8 GB
Updated	2026-06-09
Maintainer	thewordsmiths
Model Type	llama
Model Files	4.7 GB: 1-of-2 1.1 GB: 2-of-2
Supported Languages	en
Model Architecture	LlamaForCausalLM
License	apache-2.0
Context Length	8192
Model Max Length	8192
Transformers Version	4.41.2
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<\|reserved_special_token_250\|>
Vocabulary Size	128256
Torch Data Type	bfloat16

Rank the Llama SciQ 4bits 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 SciQ 4bits

Best Alternatives	Context / RAM	Downloads	Likes
Ko Pt Model Test1	8K / 16.4 GB	2027	0
Hola	8K / 5.8 GB	6	0
...ma3 Quantized 5epochs 24 06 04	8K / 5.8 GB	7	0
RAFT Llama3 Version 24.06.01	8K / 5.8 GB	5	0
Llama 3 Merged Linear	8K / 5.8 GB	16	0
Llama 3 Ko Luxia Instruct	8K / 16.4 GB	6	3
KernelLLM Unsloth Bnb 4bit	128K / 6 GB	6	2
KernelLLM Bnb 4bit	128K / 5.7 GB	9	1
Llama3.1 Merged	128K / 5.8 GB	6	0
EPFL TA Meister 4bit	8K / 5.8 GB	6	0

Note: green Score (e.g. "73.2") means that the model is better than thewordsmiths/Llama_SciQ_4bits.