LLM EXPLORER 65,695 MODELS INDEXED

SimpleSD 4B Instruct by apple

By apple · 202 downloads
HF Repo ↗ Copy URL

SimpleSD 4B Instruct is an open-source language model by apple. Features: 4b LLM, VRAM: 8.1GB, Context: 256K, License: apple-amlr, Instruction-Based, LLM Explorer Score: 0.23.

  Arxiv:2604.01193 Base model:finetune:qwen/qwen3... Base model:qwen/qwen3-4b-instr...   Code-generation   Conversational   Endpoints compatible   Instruct   Qwen3   Region:us   Safetensors   Self-distillation   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/apple/SimpleSD-4B-instruct 

SimpleSD 4B Instruct Parameters and Internals

LLM NameSimpleSD 4B Instruct
Repository 🤗https://huggingface.co/apple/SimpleSD-4B-instruct 
Base Model(s)  Qwen3 4B Instruct 2507   Qwen/Qwen3-4B-Instruct-2507
Model Size4b
Required VRAM8.1 GB
Updated2026-07-13
Maintainerapple
Model Typeqwen3
Instruction-BasedYes
Model Files  4.0 GB: 1-of-3   4.0 GB: 2-of-3   0.1 GB: 3-of-3
Model ArchitectureQwen3ForCausalLM
Licenseapple-amlr
Context Length262144
Model Max Length262144
Transformers Version4.57.3
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size151936
Errorsreplace

Best Alternatives to SimpleSD 4B Instruct

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3 4B Instruct 2507256K / 8.1 GB3109972911
FastContext 1.0 4B SFT256K / 8.1 GB5735357
Dwight 4B256K / 8.1 GB37410
SRA RiskGate 4B256K / 8 GB26621
Qwen3 4B Deslop V4256K / 8.1 GB1288
...orm3200 Impl Verl S11 Step 039256K / 8.1 GB4541
...orm3200 Impl Verl S11 Step 062256K / 8.1 GB4031
Qwen3 4B Instruct 2507 FP8256K / 5.2 GB110074478
FastContext 1.0 4B RL256K / 8.1 GB455961
NEIRA HF1 4B256K / 8 GB2880
Note: green Score (e.g. "73.2") means that the model is better than apple/SimpleSD-4B-instruct.