What are the hardware requirements for Llama3 Math Merged4?

Llama3 Math Merged4 requires approximately 22.2 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Llama3 Math Merged4 and how large is it?

Llama3 Math Merged4 is developed by Zainb114, a model with 8b 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 Llama3 Math Merged4?

Llama3 Math Merged4 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Llama3 Math Merged4 by Zainb114 — VRAM 22.2GB

Name: Llama3 Math Merged4
Author: Zainb114

Llama3 Math Merged4 is an open-source language model by Zainb114. Features: 8b LLM, VRAM: 22.2GB, Quantized, LLM Explorer Score: 0.23.

Arxiv:1910.09700 4bit Base model:adapter:unsloth/lla... Base model:unsloth/llama-3-8b-... Conversational Endpoints compatible Llama Lora Peft Quantized Region:us Safetensors Sft Sharded Tensorflow Trl Unsloth

Model Card on HF 🤗: https://huggingface.co/Zainb114/llama3-math-merged4

Llama3 Math Merged4 Parameters and Internals

LLM Name	Llama3 Math Merged4
Repository 🤗	https://huggingface.co/Zainb114/llama3-math-merged4
Base Model(s)	Llama 3 8B Bnb 4bit unsloth/llama-3-8b-bnb-4bit
Model Size	8b
Required VRAM	22.2 GB
Updated	2026-07-19
Maintainer	Zainb114
Model Files	0.3 GB 5.0 GB: 1-of-2 5.0 GB: 1-of-4 1.1 GB: 2-of-2 5.0 GB: 2-of-4 4.9 GB: 3-of-4 1.2 GB: 4-of-4
Quantization Type	4bit
Model Architecture	AutoModelForCausalLM
Model Max Length	8192
Is Biased	none
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<\|reserved_special_token_250\|>
PEFT Type	LORA
LoRA Model	Yes
PEFT Target Modules	k_proj\|o_proj\|gate_proj\|q_proj\|up_proj\|v_proj\|down_proj
LoRA Alpha	64
LoRA Dropout	0
R Param	32

Rank the Llama3 Math Merged4 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 Llama3 Math Merged4

Best Alternatives	Context / RAM	Downloads	Likes
Lora Model 1	8K / 5.8 GB	10	0
...3 Project Management Assistant	0K / 0.2 GB	4	3
LLAMA3 Vuln Detection	0K / 0.2 GB	6	0
Educate Ai V2	0K / 0.2 GB	5	0
Llama3 8B Myfine	0K / 16.1 GB	7	0
...ama 3 8B Instruct Bnb Telcom 3	0K / 0.2 GB	6	0
Model Name	0K / 16.1 GB	5	0
Llama30	0K / 16.1 GB	6	0
Llama26	0K / 16.1 GB	5	0
Llama24	0K / 16.1 GB	10	0