exolabs/qwen3-6-27b-ud-mlx-4bit-text-dequant-bf16

VISIONConcurrent Unit Cost:2Model Size:27BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Jun 3, 2026License:otherArchitecture:Transformer Featherless Exclusive Cold

The exolabs/qwen3-6-27b-ud-mlx-4bit-text-dequant-bf16 is a 27 billion parameter Qwen3.6 model, derived from the original Qwen/Qwen3.6-27B, with a context length of 32768 tokens. This specific version is a dequantized bfloat16 export of an MLX 4-bit quantized checkpoint, intended for vLLM validation. It is designed for text-only applications, providing a high-fidelity representation of the quantized model for evaluation purposes.

Loading preview...

Model Overview

The exolabs/qwen3-6-27b-ud-mlx-4bit-text-dequant-bf16 is a 27 billion parameter language model based on the Qwen3.6 architecture, originally developed by Qwen. This particular release is a specialized export, representing a dequantized bfloat16 version of an MLX 4-bit quantized checkpoint. It maintains a substantial context length of 32768 tokens, making it suitable for processing extensive textual inputs.

Key Characteristics

  • Model Family: Qwen3.6-27B, originating from Qwen/Qwen3.6-27B.
  • Parameter Count: 27 billion parameters.
  • Export Type: Dequantized bfloat16 export from an MLX 4-bit quantized source (unsloth/Qwen3.6-27B-UD-MLX-4bit).
  • Purpose: Primarily intended for vLLM validation, offering a high-precision representation of the quantized model.
  • Functionality: This is a text-only export, focusing solely on language processing tasks.

Intended Use Cases

This model is specifically designed for:

  • vLLM Validation: Evaluating the performance and fidelity of the dequantized bfloat16 model within the vLLM framework.
  • Research and Development: Analyzing the impact of dequantization on model behavior and output quality.
  • Text-based Applications: Leveraging its 27 billion parameters and large context window for various text generation, comprehension, and analysis tasks where the bfloat16 precision is desired.