Solstice-AI/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU-mlx-6Bit

feat: update tokenizer_config chat_template to Solstice 10-level solo + swarm

feat: update to Solstice 10-level solo + swarm cognitive architecture

docs: explicitly specify Mac mini M5 Pro (64GB) in performance table

docs: add restored MTP module, MLX speculative decoding, and DSpark serving recipes

fix(mtp): register 15-tensor MTP module in index and set root num_nextn_predict_layers

feat(mtp): add restored 15-tensor MTP speculative drafting module (BF16)

docs: correct empirical decode vs prefill throughput and clean MLX-LM serving recipes

docs: update chat template with <think> reasoning tokens

docs: 3x expanded model card with Apple MLX 6-bit affine specs, Anvil Metal 262K recipes, and clean sweep benchmarks

docs: add institutional benchmark-verified model card with benchmarks and MLX quickstart

Upload folder using huggingface_hub

initial commit