LLM EXPLORER 57,918 MODELS INDEXED

Step 3.5 Flash 4bit by mlx-community

By mlx-community · 1164 downloads
HF Repo ↗ Copy URL

Step 3.5 Flash 4bit is an open-source language model by mlx-community. Features: 197b LLM, VRAM: 110.3GB, Context: 256K, License: apache-2.0, Quantized, LLM Explorer Score: 0.25.

  4-bit   4bit Base model:quantized:stepfun-a... Base model:stepfun-ai/step-3.5...   Conversational   Custom code   Mlx   Quantized   Region:us   Safetensors   Sharded   Step3p5   Tensorflow
Model Card on HF 🤗: https://huggingface.co/mlx-community/Step-3.5-Flash-4bit 

Step 3.5 Flash 4bit Parameters and Internals

LLM NameStep 3.5 Flash 4bit
Repository 🤗https://huggingface.co/mlx-community/Step-3.5-Flash-4bit 
Base Model(s)  stepfun-ai/Step-3.5-Flash   stepfun-ai/Step-3.5-Flash
Model Size197b
Required VRAM110.3 GB
Updated2026-07-09
Maintainermlx-community
Model Typestep3p5
Model Files  5.1 GB: 1-of-22   5.2 GB: 2-of-22   5.2 GB: 3-of-22   5.2 GB: 4-of-22   5.2 GB: 5-of-22   5.2 GB: 6-of-22   5.2 GB: 7-of-22   5.2 GB: 8-of-22   5.2 GB: 9-of-22   5.2 GB: 10-of-22   5.2 GB: 11-of-22   5.2 GB: 12-of-22   5.2 GB: 13-of-22   5.2 GB: 14-of-22   5.2 GB: 15-of-22   5.2 GB: 16-of-22   5.2 GB: 17-of-22   5.2 GB: 18-of-22   5.2 GB: 19-of-22   5.2 GB: 20-of-22   5.2 GB: 21-of-22   1.2 GB: 22-of-22
Quantization Type4bit
Model ArchitectureStep3p5ForCausalLM
Licenseapache-2.0
Context Length262144
Model Max Length262144
Tokenizer ClassTokenizersBackend
Padding Token<｜end▁of▁sentence｜>
Vocabulary Size128896
Torch Data Typebfloat16

Best Alternatives to Step 3.5 Flash 4bit

Best Alternatives
Context / RAM
Downloads
Likes
Step 3.5 Flash 8bit256K / 210.7 GB2632
Step 3.5 Flash Bf16256K / 138.5 GB481
Note: green Score (e.g. "73.2") means that the model is better than mlx-community/Step-3.5-Flash-4bit.