LLM EXPLORER 58,608 MODELS INDEXED

SERA 32B 4bit by mlx-community

By mlx-community · 12 downloads
HF Repo ↗ Copy URL

SERA 32B 4bit is an open-source language model by mlx-community. Features: 32b LLM, VRAM: 18.5GB, Context: 40K, License: apache-2.0, Quantized, LLM Explorer Score: 0.22.

  4-bit   4bit   Base model:allenai/sera-32b Base model:quantized:allenai/s...   Conversational   Mlx   Quantized   Qwen3   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/mlx-community/SERA-32B-4bit 

SERA 32B 4bit Parameters and Internals

LLM NameSERA 32B 4bit
Repository 🤗https://huggingface.co/mlx-community/SERA-32B-4bit 
Base Model(s)  SERA 32B   allenai/SERA-32B
Model Size32b
Required VRAM18.5 GB
Updated2026-08-10
Maintainermlx-community
Model Typeqwen3
Model Files  5.4 GB: 1-of-4   5.3 GB: 2-of-4   5.4 GB: 3-of-4   2.4 GB: 4-of-4
Quantization Type4bit
Model ArchitectureQwen3ForCausalLM
Licenseapache-2.0
Context Length40960
Model Max Length40960
Transformers Version4.57.1
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size151936
Errorsreplace

Best Alternatives to SERA 32B 4bit

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3 32B Bnb 4bit40K / 19.2 GB771257
... E5 Grad Accum 16 Merged 16bit40K / 65.8 GB260
...llcpt E5 Baseline Merged 16bit40K / 65.8 GB230
...line Merged 16bit Merged 16bit40K / 65.8 GB200
SERA 32B 6bit40K / 26.6 GB272
SERA 32B GA 4bit40K / 18.5 GB71
Qwen3 32B Unsloth Bnb 4bit40K / 39.5 GB667415
Qwen3 32B 4bit40K / 18.5 GB24366
Qwen3 32B MLX 8bit40K / 33.7 GB74912
Qwen3 32B MLX 4bit40K / 17.4 GB5738
Note: green Score (e.g. "73.2") means that the model is better than mlx-community/SERA-32B-4bit.