bartowski/ArliAI_Qwen3.5-27B-RpRMax-v1-GGUF

Same weights, converted vs ArliAI/Qwen3.5-27B-RpRMax-v1. Same weights stored in another dtype (BF16/F32 → BF16/F32), plus 48 new tensors.

Tensors changed: lm_head 0/1, Embeddings 0/1, MLP 0/192, Attention 0/64, Norms 0/161

On the Hub · Downloads on Model Pulse