Unsloth has released new Qwen3.8-27B GGUFs utilizing the Dynamic v3.0 quantization method. This update achieves a 10% increase in accuracy across benchmarks like Div-300 and KLD compared to previous versions. Additionally, new 1-bit quants have been introduced, retaining 77% accuracy while remaining functional on 8GB of RAM.