LLM EXPLORER 61,405 MODELS INDEXED

Apertus V1.1 4B Instruct MLX INT3 by swiss-ai

By swiss-ai · 53 downloads
HF Repo ↗ Copy URL

Apertus V1.1 4B Instruct MLX INT3 is an open-source language model by swiss-ai. Features: 4b LLM, VRAM: 2.2GB, Context: 4K, License: apache-2.0, Instruction-Based, LLM Explorer Score: 0.24.

  Arxiv:2509.14233   Arxiv:2605.29128   3-bit   Apertus Base model:quantized:swiss-ai/... Base model:swiss-ai/apertus-v1...   Compliant   Conversational   Endpoints compatible   Instruct   Multilingual   Region:us   Safetensors   Swiss-ai
Model Card on HF 🤗: https://huggingface.co/swiss-ai/Apertus-v1.1-4B-Instruct-MLX-INT3 

Apertus V1.1 4B Instruct MLX INT3 Parameters and Internals

LLM NameApertus V1.1 4B Instruct MLX INT3
Repository 🤗https://huggingface.co/swiss-ai/Apertus-v1.1-4B-Instruct-MLX-INT3 
Base Model(s)  swiss-ai/Apertus-v1.1-4B-Instruct   swiss-ai/Apertus-v1.1-4B-Instruct
Model Size4b
Required VRAM2.2 GB
Updated2026-08-09
Maintainerswiss-ai
Model Typeapertus
Instruction-BasedYes
Model Files  2.2 GB
Model ArchitectureApertusForCausalLM
Licenseapache-2.0
Context Length4096
Model Max Length4096
Transformers Version5.5.0
Tokenizer ClassTokenizersBackend
Padding Token<pad>
Vocabulary Size131072

Best Alternatives to Apertus V1.1 4B Instruct MLX INT3

Best Alternatives
Context / RAM
Downloads
Likes
Apertus V1.1 4B Instruct4K / 7.7 GB11032
...rtus V1.1 4B Instruct MLX INT44K / 2.6 GB801
...rtus V1.1 4B Instruct MLX INT64K / 3.4 GB561
Note: green Score (e.g. "73.2") means that the model is better than swiss-ai/Apertus-v1.1-4B-Instruct-MLX-INT3.