BRAVO!
Just wanted to say that you continue to be the GOAT for me when it comes to GGUFs for creative writing. This one really hits it out of the park. GLM-4.6 is also great, but I think I might even like 4.5 a little better π€·ββοΈ
Even my small IQ2 quant of 4.5 is doing just as good, if not better than the current late 2025 big frontier models. I was blown away with its ability to comprehend and carry out complex instructions (e.g., story building, world building, etc.) EXACTLY as I envisioned. A lot of other models seem to get lost in the sauce and/or default to its own tendencies at some point. But GLM-4.5 somehow "listens (haha)" very well in terms of what I want and in terms of what I'm going for.
Unsloth also does a great job with GGUF quantization as well, but tbh I def gotta give you the edge in terms of quants that are best for creative writing.