Ministral 3 8B Instruct 2512 Bnb 4bit is an open-source language model by unsloth. Features: 8b LLM, VRAM: 6.8GB, License: apache-2.0, Quantized, Instruction-Based, LLM Explorer Score: 0.24.
4-bit 4bit Ar Base model:mistralai/ministral... Base model:quantized:mistralai... Bitsandbytes De En Es Fr Instruct It Ja Ko Mistral Mistral-common Mistral3 Nl Pt Quantized Region:us Safetensors Unsloth Vllm Zh
Model Card on HF 🤗: https://huggingface.co/unsloth/Ministral-3-8B-Instruct-2512-bnb-4bit
Ministral 3 8B Instruct 2512 Bnb 4bit Benchmarks
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
What is the LLM Explorer Rank (Score)
Ministral 3 8B Instruct 2512 Bnb 4bit Parameters and Internals
| LLM Name | Ministral 3 8B Instruct 2512 Bnb 4bit |
| Repository 🤗 | https://huggingface.co/unsloth/Ministral-3-8B-Instruct-2512-bnb-4bit |
| Base Model(s) | |
| Model Size | 8b |
| Required VRAM | 6.8 GB |
| Updated | 2026-06-12 |
| Maintainer | unsloth |
| Model Type | mistral3 |
| Instruction-Based | Yes |
| Model Files | |
| Supported Languages | en fr es de it pt nl zh ja ko ar |
| Quantization Type | 4bit |
| Model Architecture | Mistral3ForConditionalGeneration |
| License | apache-2.0 |
| Model Max Length | 262144 |
| Transformers Version | 5.0.0.dev0 |
| Tokenizer Class | LlamaTokenizer |
| Padding Token | <pad> |
| Torch Data Type | bfloat16 |
Best Alternatives to Ministral 3 8B Instruct 2512 Bnb 4bit
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| Ministral 3 8B Instruct 2512 | 0K / 16.9 GB | 264 | 2 |
Rank the Ministral 3 8B Instruct 2512 Bnb 4bit Capabilities
🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟
What open-source LLMs or SLMs are you in search of? 54415 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a