Instructions to use Youssofal/Qwen3.6-27B-MTPLX-Optimized-Speed-FP16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use Youssofal/Qwen3.6-27B-MTPLX-Optimized-Speed-FP16 with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir Qwen3.6-27B-MTPLX-Optimized-Speed-FP16 Youssofal/Qwen3.6-27B-MTPLX-Optimized-Speed-FP16
- Notebooks
- Google Colab
- Kaggle
- Local Apps
- LM Studio
| { | |
| "arch_id": "qwen3-next-mtp", | |
| "artifact_role": "maximum-speed-flat4-candidate", | |
| "base_trunk": "mlx-community/Qwen3.6-27B-4bit", | |
| "exactness_baseline": { | |
| "attention_impl": "mlx_vector_paged", | |
| "context": 64, | |
| "gate": "phase0h-paged-verifier-exactness", | |
| "max_abs_diff": 0.0, | |
| "mode": "decode-from-stock-prefix", | |
| "sample_agreement": 1.0, | |
| "status": "passed", | |
| "topk_overlap_ratio": 1.0, | |
| "total_variation": 0.0, | |
| "verify_tokens": 4 | |
| }, | |
| "mtp_depth_max": 3, | |
| "mtp_sidecar": "Qwen3.6-27B-MTPLX-CyanKiwi-Packed-BF16-INT4-v3", | |
| "mtplx_version": "0.1.0-preview", | |
| "precision_policy": { | |
| "intended_default_for": [ | |
| "m1", | |
| "m2" | |
| ], | |
| "note": "This is a sibling precision variant; it is not a universal speed claim.", | |
| "routing": "mtplx start auto-selects this artifact on M1/M2 Apple Silicon", | |
| "source_repo": "Youssofal/Qwen3.6-27B-MTPLX-Optimized-Speed", | |
| "variant": "fp16" | |
| }, | |
| "precision_variant": "fp16", | |
| "recommended_draft_lm_head": { | |
| "bits": 3, | |
| "group_size": 64, | |
| "mode": "affine" | |
| }, | |
| "recommended_draft_sampler": { | |
| "temperature": 0.7, | |
| "top_k": 20, | |
| "top_p": 0.95 | |
| }, | |
| "recommended_profile": "performance-cold", | |
| "sampler": { | |
| "temperature": 0.6, | |
| "top_k": 20, | |
| "top_p": 0.95 | |
| }, | |
| "speed_evidence": { | |
| "acceptance_by_depth": [ | |
| 1.0, | |
| 0.9795918367346939, | |
| 0.9387755102040817 | |
| ], | |
| "accepted_by_depth": [ | |
| 49, | |
| 48, | |
| 46 | |
| ], | |
| "artifacts": [ | |
| "/Users/youssof/.mtplx/fanmax_flat4_draft_temp070_cleanapps_20260503-230229.json", | |
| "/Users/youssof/.mtplx/fanmax_flat4_draft_temp070_cleanapps_repeat_20260503-230458.json" | |
| ], | |
| "correction_tokens": 3, | |
| "depth": 3, | |
| "draft_lm_head": "3-bit affine group64 draft-only head", | |
| "draft_sampler": { | |
| "temperature": 0.7, | |
| "top_k": 20, | |
| "top_p": 0.95 | |
| }, | |
| "drafted_by_depth": [ | |
| 49, | |
| 49, | |
| 49 | |
| ], | |
| "enable_thinking": false, | |
| "greedy_diagnostic": { | |
| "accepted_by_depth": [ | |
| 49, | |
| 49, | |
| 43 | |
| ], | |
| "artifact": "/Users/youssof/.mtplx/fanmax_flat4_full_greedy_cleanapps_20260503-230229.json", | |
| "correction_tokens": 0, | |
| "draft_sampler": { | |
| "temperature": 0.0, | |
| "top_k": 0, | |
| "top_p": 1.0 | |
| }, | |
| "drafted_by_depth": [ | |
| 51, | |
| 51, | |
| 51 | |
| ], | |
| "target_sampler": { | |
| "temperature": 0.0, | |
| "top_k": 0, | |
| "top_p": 1.0 | |
| }, | |
| "tok_s": 60.108346853627076, | |
| "verify_calls": 51 | |
| }, | |
| "older_3bit_artifacts": [ | |
| "/Users/youssof/.mtplx/fanmax_flat4_cyankiwi_draft_head_3bit_isolated_20260503.json", | |
| "/Users/youssof/.mtplx/fanmax_flat4_cyankiwi_draft_head_3bit_repeat_20260503.json" | |
| ], | |
| "older_3bit_tok_s": [ | |
| 60.03822277200243, | |
| 60.06063543439426 | |
| ], | |
| "previous_best_artifacts": [ | |
| "/Users/youssof/.mtplx/fanmax_flat4_draft_temp070_optimized_env_20260503.json", | |
| "/Users/youssof/.mtplx/fanmax_flat4_draft_temp070_repeat_optimized_env_20260503.json" | |
| ], | |
| "previous_best_tok_s": [ | |
| 61.52744855840259, | |
| 60.89966247416089 | |
| ], | |
| "production_cli_contract_checks": [ | |
| { | |
| "artifact": "/Users/youssof/.mtplx/fanmax_mtplx_flat4_cyankiwi_cli_contract_perfcold_disablethinking_20260503.json", | |
| "tok_s": 59.51602930081416 | |
| }, | |
| { | |
| "artifact": "/Users/youssof/.mtplx/fanmax_mtplx_flat4_cyankiwi_cli_contract_perfcold_disablethinking_repeat_20260503.json", | |
| "tok_s": 59.97165246907193 | |
| } | |
| ], | |
| "profile": "performance-cold", | |
| "timestamp": "2026-05-03T23:07:00+0100", | |
| "tok_s": [ | |
| 63.05651084048216, | |
| 62.885564723941826 | |
| ], | |
| "verify_calls": 49 | |
| }, | |
| "verified_on": { | |
| "hardware": "Apple M5 Max, 128 GB unified memory", | |
| "machine_arch": "arm64", | |
| "model": "Qwen3.6-27B-MTPLX-Flat4-CyanKiwiMTP", | |
| "timestamp": "2026-05-03T17:49:00+0100" | |
| } | |
| } | |