LLM EXPLORER 66,143 MODELS INDEXED

Synapse Alpha 1.0 4bit by evraldtah

By evraldtah · 43 downloads
HF Repo ↗ Copy URL

Synapse Alpha 1.0 4bit is an open-source language model by evraldtah. Features: 2b LLM, VRAM: 1.1GB, License: apache-2.0, Quantized, LLM Explorer Score: 0.29.

  4-bit   4bit   Agent Base model:quantized:qwen/qwen...   Base model:qwen/qwen3.5-2b   Conversational   Mlx   Notes   On-device   Quantized   Qwen3 5   Region:us   Safetensors   Summarization   Synapse   Transcription
Model Card on HF 🤗: https://huggingface.co/evraldtah/Synapse-Alpha-1.0-4bit 

Synapse Alpha 1.0 4bit Parameters and Internals

LLM NameSynapse Alpha 1.0 4bit
Repository 🤗https://huggingface.co/evraldtah/Synapse-Alpha-1.0-4bit 
Base Model(s)  Qwen/Qwen3.5-2B   Qwen/Qwen3.5-2B
Model Size2b
Required VRAM1.1 GB
Updated2026-10-30
Maintainerevraldtah
Model Typeqwen3_5
Model Files  1.1 GB
Quantization Type4bit
Model ArchitectureQwen3_5ForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Transformers Version4.57.0.dev0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Errorsreplace

Best Alternatives to Synapse Alpha 1.0 4bit

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3.5 2B M40 6bit0K / 1.5 GB4631
Qwen3.5 2B Base 6bit0K / 1.5 GB4171
Qwen3.5 2B 4bit Slim0K / 1.1 GB2271
...Bonsai 2 2B ONNX WebGPU Vision0K /  GB1881
Loopl 2B 4bit0K / 1.7 GB1371
Qwen3.5 2B MLX 4bit0K / 1.7 GB123102
Qwen3.8 2B Distill Mlx 5Bit0K / 1.3 GB3871
Qwen3.5 2B 6bit0K / 2.2 GB86807
Qwen3.5 2B 4bit0K / 1.7 GB53745
Mavery Geselle 2B MLX 4bit0K / 1.7 GB772
Note: green Score (e.g. "73.2") means that the model is better than evraldtah/Synapse-Alpha-1.0-4bit.