---
title: SynLogic-Mix-3-32B-mlx-OptiQ-5bit
canonical_url: "https://www.modelscope.cn/models/latentnode/SynLogic-Mix-3-32B-mlx-OptiQ-5bit"
md_url: "https://www.modelscope.cn/models/latentnode/SynLogic-Mix-3-32B-mlx-OptiQ-5bit.md"
repository: latentnode/SynLogic-Mix-3-32B-mlx-OptiQ-5bit
chinese_name: SynLogic-Mix-3-32B-mlx-OptiQ-5bit
last_updated: 2026-07-27
license: mit
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - qwen2
architectures:
  - Qwen2ForCausalLM
base_model:
  - MiniMaxAI/SynLogic-Mix-3-32B
base_model_relation: quantized
parameters: 6.1B
tensor_type:
  - BF16
  - U32
library_name:
  - mlx
  - safetensors
language:
  - en
downloads: 8
stars: 0
tags:
  - LLM
  - mlx
---

# SynLogic-Mix-3-32B-mlx-OptiQ-5bit

> SynLogic-Mix-3-32B-mlx-OptiQ-5bit - latentnode 在 ModelScope 开源的模型。Yet another showcase that mlx-optiq is more than the developers claim: \ It is not limited to Qwen3.5-3.6, but also can be used for all qwen family, and the same seems to go for Gemma – not…

latentnode/SynLogic-Mix-3-32B-mlx-OptiQ-5bit 是 ModelScope 魔搭社区上的 6.1B 参数text-generation模型，采用 mit 许可，基于 MiniMaxAI/SynLogic-Mix-3-32B 构建。

- **Repository**: latentnode/SynLogic-Mix-3-32B-mlx-OptiQ-5bit
- **License**: mit
- **Tasks**: text-generation
- **Parameters**: 6.1B
- **Base model**: MiniMaxAI/SynLogic-Mix-3-32B
- **Tags**: LLM, mlx
- **Downloads**: 8
- **Stars**: 0
- **Last updated**: 2026-07-27

Source: https://www.modelscope.cn/models/latentnode/SynLogic-Mix-3-32B-mlx-OptiQ-5bit

---

Yet another showcase that mlx-optiq is more than the developers claim: \
It is not limited to Qwen3.5-3.6, but also can be used for all qwen family, and the same seems to go for Gemma – not just gemma4, but 3 as well (at least)! \
All that extends to other models built on their top.