What are the hardware requirements for Luca MN Bf16?

Luca MN Bf16 requires approximately 24.5 GB of VRAM and supports a context window of 128K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Luca MN Bf16 and how large is it?

Luca MN Bf16 is developed by rAIfle, a model with 12.2b 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 Luca MN Bf16?

Luca MN Bf16 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Luca MN Bf16 by rAIfle — VRAM 24.5GB, 128K context

Name: Luca MN Bf16
Author: rAIfle

Luca MN Bf16 is an open-source language model by rAIfle. Features: 12.2b LLM, VRAM: 24.5GB, Context: 128K, License: apache-2.0, Quantized, LLM Explorer Score: 0.14.

4bit Base model:finetune:unsloth/mi... Base model:unsloth/mistral-nem... Conversational Endpoints compatible Mistral Quantized Region:us Safetensors Sft Sharded Tensorflow Trl Unsloth

Model Card on HF 🤗: https://huggingface.co/rAIfle/Luca-MN-bf16

Luca MN Bf16 Parameters and Internals

Training Details

Data Sources:

some RP data, c2-data, jondurbin/gutenberg-dpo-v0.1

Methodology:

High-r LoRA-pass over Nemo-Base with 2 epochs of some RP data, then a low-r pass with 0.5 epochs of the c2-data, then 3 epochs of DPO using jondurbin/gutenberg-dpo-v0.1.

Input Output

Input Format:

Use the Mistral V3-Tekken context- and instruct-templates.

Performance Tips:

Temperature at about 1.25 seems to be the sweet spot, with either MinP at 0.05 or TopP at 0.9. DRY/Smoothing etc depending on your preference.

LLM Name	Luca MN Bf16
Repository 🤗	https://huggingface.co/rAIfle/Luca-MN-bf16
Base Model(s)	...istral Nemo Base 2407 Bnb 4bit unsloth/Mistral-Nemo-Base-2407-bnb-4bit
Model Size	12.2b
Required VRAM	24.5 GB
Updated	2026-08-02
Maintainer	rAIfle
Model Type	mistral
Model Files	4.9 GB: 1-of-5 4.9 GB: 2-of-5 4.9 GB: 3-of-5 4.9 GB: 4-of-5 4.9 GB: 5-of-5
Quantization Type	4bit
Model Architecture	MistralForCausalLM
License	apache-2.0
Context Length	131072
Model Max Length	131072
Transformers Version	4.44.2
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<pad>
Vocabulary Size	131072
Torch Data Type	bfloat16

Rank the Luca MN Bf16 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 Luca MN Bf16

Best Alternatives	Context / RAM	Downloads	Likes
NemoR	1000K / 24.5 GB	24	0
Nemo Carpmuscle V0.1	1000K / 24.5 GB	4	1
...d 1.0 Nemo Base 2407 Ita 16bit	1000K / 24.5 GB	2579	4
NaturalLM	1000K / 24.5 GB	163	6
NaturalLM	1000K / 24.5 GB	27	6
... Mistral Nemo Base 2407 Sft V1	1000K / 24.5 GB	49	1
...all MathMistral Nemo Base 2407	1000K / 24.5 GB	13	1
RAISED Mistral Nemo SFT	128K / 24.5 GB	12	0
...aptain Eris Violet GRPO V0.420	1000K / 24.5 GB	23	24
...istral Nemo Instruct 2407 MPOA	1000K / 24.5 GB	37	3

Note: green Score (e.g. "73.2") means that the model is better than rAIfle/Luca-MN-bf16.