LLM EXPLORER 61,755 MODELS INDEXED

Bonsai 2 27B Ternary JANG by OsaurusAI

By OsaurusAI · 978 downloads
HF Repo ↗ Copy URL

Bonsai 2 27B Ternary JANG is an open-source language model by OsaurusAI. Features: 27b LLM, VRAM: 8.1GB, License: apache-2.0, Quantized, LLM Explorer Score: 0.34.

  2bit   Affine-quantization Base model:finetune:prism-ml/t... Base model:prism-ml/ternary-bo...   Bonsai   Bonsai-2   Conversational   Hadamard   Image-text-to-text   Jang   Mlx   Multimodal   Quantized   Qwen3.8   Qwen3 5   Reasoning   Region:us   Safetensors   Sharded   Tensorflow   Ternary   Thinking   Video   Vision
Model Card on HF 🤗: https://huggingface.co/OsaurusAI/Bonsai-2-27B-Ternary-JANG 

Bonsai 2 27B Ternary JANG Parameters and Internals

LLM NameBonsai 2 27B Ternary JANG
Repository 🤗https://huggingface.co/OsaurusAI/Bonsai-2-27B-Ternary-JANG 
Base Model(s)  Ternary Bonsai 2 27B Mlx 2bit   prism-ml/Ternary-Bonsai-2-27B-mlx-2bit
Model Size27b
Required VRAM8.1 GB
Updated2026-09-20
MaintainerOsaurusAI
Model Typeqwen3_5
Model Files  2.1 GB: 1-of-4   2.1 GB: 2-of-4   2.1 GB: 3-of-4   1.8 GB: 4-of-4
Quantization Type2bit
Model ArchitectureQwen3_5ForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Errorsreplace

Best Alternatives to Bonsai 2 27B Ternary JANG

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3.6 27B OptiQ 4bit256K / 18.7 GB1306559
Bonsai 27B Mlx 1bit0K / 5.1 GB809309205
Ternary Bonsai 27B Mlx 2bit0K / 8.5 GB809024168
Qwen3.8 27B MLX 5bit0K / 19.5 GB46750442
Qwen3.6 27B MLX 8bit0K / 29.4 GB87914514
Qwen3.6 27B MLX 6bit0K / 22.7 GB8129772
Qwen3.8 27B 8bit0K / 29.4 GB1389233
Qwen3.6 27B MLX 4bit0K / 16 GB84686116
...wift Qwen3.8 27B OQ4e Fp16 Mtp0K / 17.8 GB66310
Qwen3.8 27B MLX 8bit0K / 29.4 GB248117
Note: green Score (e.g. "73.2") means that the model is better than OsaurusAI/Bonsai-2-27B-Ternary-JANG.