Kwen-4B (Q4_K_M GGUF)
This is a fine-tuned version of Qwen3-4B, customized by the Kwen Foundation.
Highlights:
- Identity Baked: Hard-coded to identify as "Kwen" across 25 training epochs.
- Reasoning Enabled: Maintains the original Qwen3 "Thinking" architecture while prioritizing the Kwen identity.
- Optimized for Local: Quantized to Q4_K_M for peak performance on consumer GPUs like the RTX 4060.
Usage:
This model works best in Interactive Mode with thinking enabled.
