tarruda releases Step 3.7 Flash IQ4_XS GGUF with preserve_thinking

User tarruda has uploaded a new GGUF model file for the Step-3.7 Flash architecture to Hugging Face. The release specifically includes an IQ4_XS quantization variant that incorporates the preserve_thinking feature.

• Model: Step 3.7 Flash
• Format: GGUF
• Quantization: IQ4_XS
• Feature: preserve_thinking enabled