Qwen3.8-Whittle-MoE-27B-MLX-4bit

Description

Qwen3.8-Whittle-MoE-27B-MLX-4bit is an open-weight text-generation model with roughly 27B parameters published by barozp on the Hugging Face Hub. It has 1,470 downloads. Runs with mlx-lm.

Categories

• Research
• Productivity

Tags

• mlx-lm
• qwen3_5_moe
• qwen
• qwen3
• reasoning
• opus-distill
• whittle
• expert-pruning

Capabilities

• Text generation
• Conversational
• Library: mlx-lm
• 27B parameters