OrcaSAQ-2 27B

OrcaSAQ-2 27B is a highly compressed, sensitivity-aware mixed-precision quantized version of the Qwen3.8-27B model. It reduces the storage footprint by over 77% while maintaining high fidelity to the original model, making it suitable for deployment on hardware with limited GPU memory.

OrcaSAQ-2 27B is a highly compressed, sensitivity-aware mixed-precision quantized version of the Qwen3.8-27B model. It reduces the storage footprint by over 77% while maintaining high fidelity to the original model, making it suitable for deployment on hardware with limited GPU memory.

• 013-bit sensitivity-aware mixed-precision quantization
• 02Supports MTP speculative decoding and tool calling
• 03Optimized for long-horizon agents and complex reasoning
• 04262K context window support
• 05Compatible with vLLM for production serving