LLM EXPLORER 65,418 MODELS INDEXED

DeepSeek V4 Pro Qwen3.5 9B 8bit by mlx-community

By mlx-community · 241 downloads
HF Repo ↗ Copy URL

DeepSeek V4 Pro Qwen3.5 9B 8bit is an open-source language model by mlx-community. Features: 9b LLM, VRAM: 10.4GB, License: apache-2.0, Quantized, LLM Explorer Score: 0.26.

  8-bit   8bit   Agent Base model:jackrong/deepseek-v... Base model:quantized:jackrong/...   Coder   Conversational   En   Es   Function-calling   Image-text-to-text   Ja   Long-context   Mlx   Moe   Multimodal   Quantized   Qwen3 5   Qwen3 5 moe   Qwen3 6   Region:us   Ru   Safetensors   Sharded   Tensorflow   Tool-use   Video   Vision   Zh
Model Card on HF 🤗: https://huggingface.co/mlx-community/DeepSeek-V4-Pro-Qwen3.5-9B-8bit 

DeepSeek V4 Pro Qwen3.5 9B 8bit Parameters and Internals

LLM NameDeepSeek V4 Pro Qwen3.5 9B 8bit
Repository 🤗https://huggingface.co/mlx-community/DeepSeek-V4-Pro-Qwen3.5-9B-8bit 
Model Nameunsloth/Qwen3.5-9B
Base Model(s)  Jackrong/DeepSeek-V4-Pro-Qwen3.5-9B   Jackrong/DeepSeek-V4-Pro-Qwen3.5-9B
Model Size9b
Required VRAM10.4 GB
Updated2026-10-22
Maintainermlx-community
Model Typeqwen3_5
Model Files  5.3 GB: 1-of-2   5.1 GB: 2-of-2
Supported Languagesen zh es ru ja
Quantization Type8bit
Model ArchitectureQwen3_5ForConditionalGeneration
Licenseapache-2.0
Model Max Length262144
Tokenizer ClassQwen2Tokenizer
Padding Token<|vision_pad|>
Errorsreplace

Best Alternatives to DeepSeek V4 Pro Qwen3.5 9B 8bit

Best Alternatives
Context / RAM
Downloads
Likes
Qwen3.5 9B MLX 4bit0K / 5.9 GB5800124
Qwen3.5 9B MLX 8bit0K / 10.4 GB5504971
...2.6 Distill Qwen 9B OptiQ 4bit0K / 7.1 GB19166
Ornith 1.5 9B MLX 4bit0K / 5 GB304318
Qwen3.5 9B MLX 4bit0K / 5.9 GB22827151
Ornith 1.0 9B 4bit0K / 5.9 GB97919
Ornith 1.5 9B MLX 8bit0K / 9.5 GB26867
Qwen3.5 9B 8bit0K / 10.4 GB4565217
... V2.6 Distill Qwen 9B MLX 4bit0K / 5.9 GB4062
LensVLM 9B OptiQ 4bit0K / 7.1 GB2425
Note: green Score (e.g. "73.2") means that the model is better than mlx-community/DeepSeek-V4-Pro-Qwen3.5-9B-8bit.