What are the hardware requirements for OpenBezoar SFT?

OpenBezoar SFT requires approximately 13.7 GB of VRAM and supports a context window of 2K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed OpenBezoar SFT and how large is it?

OpenBezoar SFT is developed by SurgeGlobal, a model with 3b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does OpenBezoar SFT perform on standard benchmarks?

OpenBezoar SFT has the following published scores: MMLU 28.46. Compare against reference models on this page or on the LLM Explorer leaderboards.

OpenBezoar SFT by SurgeGlobal — VRAM 13.7GB, 2K context

Name: OpenBezoar SFT
Author: SurgeGlobal

OpenBezoar SFT is an open-source language model by SurgeGlobal. Features: 3b LLM, VRAM: 13.7GB, Context: 2K, License: cc-by-nc-4.0, Instruction-Based, LLM Explorer Score: 0.13, Arc: 40.9, HellaSwag: 71.2, MMLU: 28.5, GSM8K: 2.5.

Arxiv:2306.02707 Arxiv:2404.12195 Base model:finetune:openlm-res... Base model:openlm-research/ope... Dataset:surgeglobal/evol-instr... Dataset:surgeglobal/lamini Dataset:surgeglobal/orca En Endpoints compatible Instruct Llama Pytorch Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/SurgeGlobal/OpenBezoar-SFT

OpenBezoar SFT Parameters and Internals

Model Type

instruction-following, text generation

Additional Notes

The model uses Q-LoRA with a configuration of r: 16, alpha: 16, dropout: 0.05 on target modules [q_proj, v_proj, k_proj]. Uses datasets LaMini, Orca, Evol-Instruct for instruction tuning.

Supported Languages

en (full)

Input Output

Input Format:

Modified Alpaca prompt template

Accepted Modalities:

text

Output Format:

Text

Performance Tips:

Use the prescribed prompt template for optimal results

LLM Name	OpenBezoar SFT
Repository 🤗	https://huggingface.co/SurgeGlobal/OpenBezoar-SFT
Base Model(s)	Open Llama 3b V2 openlm-research/open_llama_3b_v2
Model Size	3b
Required VRAM	13.7 GB
Updated	2026-06-28
Maintainer	SurgeGlobal
Model Type	llama
Instruction-Based	Yes
Model Files	10.0 GB: 1-of-2 3.7 GB: 2-of-2 10.0 GB: 1-of-2 3.7 GB: 2-of-2
Supported Languages	en
Model Architecture	LlamaForCausalLM
License	cc-by-nc-4.0
Context Length	2048
Model Max Length	2048
Transformers Version	4.33.0
Tokenizer Class	LlamaTokenizer
Vocabulary Size	32000
Torch Data Type	float32

Rank the OpenBezoar SFT 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 OpenBezoar SFT

Best Alternatives	Context / RAM	Downloads	Likes
Llama 3.2 3B Instruct	128K / 6.5 GB	1829572	2368
Llama 3.2 3B Instruct	128K / 6.5 GB	186207	96
...3B TechWriter Instruct Unsloth	128K / 6.5 GB	6005	1
...ini Math345 GroupB Llama32 End	128K / 6.5 GB	755	0
...ama 3.2 3B TechWriter Instruct	128K / 6.4 GB	226	0
Llama 3.2 3B Instruct	128K / 6.5 GB	991	9
HyLo Llama 6MLA22GDN 8K SFT	128K / 17.1 GB	42	1
Alloma 3B Instruct	128K / 6.5 GB	55456	6
Llama3.2 3B Manumit V1	128K / 6.4 GB	135	0
... V7 SFT 10voicebot Loopcleaned	128K / 6.5 GB	229	0

Note: green Score (e.g. "73.2") means that the model is better than SurgeGlobal/OpenBezoar-SFT.