LLM EXPLORER 58,425 MODELS INDEXED

M3.2 24B Loki V1.3 by CrucibleLab

By CrucibleLab · 21 downloads
HF Repo ↗ Copy URL

M3.2 24B Loki V1.3 is an open-source language model by CrucibleLab. Features: 24b LLM, VRAM: 47.3GB, Context: 128K, License: apache-2.0, LLM Explorer Score: 0.2.

  Mistral   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/CrucibleLab/M3.2-24B-Loki-V1.3 

M3.2 24B Loki V1.3 Parameters and Internals

LLM NameM3.2 24B Loki V1.3
Repository 🤗https://huggingface.co/CrucibleLab/M3.2-24B-Loki-V1.3 
Base Model(s)  mistralai/Mistral-Small-3.2-24B-Instruct-2506   mistralai/Mistral-Small-3.2-24B-Instruct-2506
Model Size24b
Required VRAM47.3 GB
Updated2026-08-14
MaintainerCrucibleLab
Model Typemistral
Model Files  2.7 GB: 1-of-10   4.8 GB: 2-of-10   4.9 GB: 3-of-10   4.8 GB: 4-of-10   4.8 GB: 5-of-10   4.9 GB: 6-of-10   4.8 GB: 7-of-10   4.8 GB: 8-of-10   4.9 GB: 9-of-10   3.2 GB: 10-of-10   2.7 GB: 11-of-11
Model ArchitectureMistralForCausalLM
Licenseapache-2.0
Context Length131072
Model Max Length131072
Transformers Version4.54.1
Tokenizer ClassLlamaTokenizerFast
Padding Token<pad>
Vocabulary Size131072
Torch Data Typebfloat16

Best Alternatives to M3.2 24B Loki V1.3

Best Alternatives
Context / RAM
Downloads
Likes
...ral Nemo Instruct 24B Merge V11000K / 24.6 GB90
...all 3.1 24B Instruct 2503 Text128K / 47.3 GB180
Morbid Miasma 24B128K / 47.3 GB21414
Zepar 24B V1128K / 47.3 GB1815
MagiSeek Pro V1128K / 47.3 GB3031
Slimaki Tavern 24B V1.3128K / 47.3 GB34616
Cydonia 24B V4.3128K / 47.3 GB3652138
Transformed Journey 24B128K / 47.3 GB650
Magistry Painted Cydonia 24B128K / 47.3 GB602
Morax 24B V2128K / 47.3 GB316
Note: green Score (e.g. "73.2") means that the model is better than CrucibleLab/M3.2-24B-Loki-V1.3.