splats/Qwen3.6-35B-A3B-oQ6e

Qwen3.6-35B-A3B — oQe Series

This repository contains Enhanced (oQe) MLX quants for the original Qwen3.6-35B-A3B. While standard quants are often made in one "streaming" pass, the oQe series uses a multi-stage optimization path to make sure the model doesn't lose its "intelligence" at lower bitrates.

🚀 The oQe (Enhanced) Build Path

We move away from simple rounding and use a more deliberate process to protect the model's logic:

1. Sensitivity Mapping: We don't guess which layers are important. We run a calibration pass to measure exactly how much precision each layer needs to keep its output stable.
2. Hessian-Based Tuning: Every time we round a weight to a lower bit, we adjust the surrounding weights to compensate for the error. This keeps the model's internal math from "drifting" as it gets smaller.
3. Unified Batching: For this MoE architecture, we process all 256 experts together so the 8+1 routing path stays stable and precise.

📋 oQ Build Performance Matrix

Tier	Target bpw	Actual bpw	Size	Precision Boosts	Hybrid Plan / Strategy
oQ8e	8.0	8.00	35.1 GB	0	Full 8-bit Static
oQ6e	6.0	6.60	27.4 GB	162	8bit×162 (DeltaNet Anchor)
oQ5e	5.0	5.67	23.6 GB	352	8bit×162, 6bit×190
oQ4e	4.0	4.70	19.7 GB	317	8bit×162, 6bit×47, 5bit×108
oQ3.5e	3.5	4.00	16.9 GB	60	8bit×10, 5bit×10, 4bit×40

🛠 Technical Build Audit

• Calibration: Uses a 600-sample dataset across code, reasoning, and multi-turn conversations.
• Sensitivity Proxy: Qwen3.6-35B-A3B-oQ8e (Internal Baseline).
• Optimization Floor: Mandatory 8-bit protection is applied to the lm_head, DeltaNet heads, and the first 6 layers to keep the "Agentic Coding" features native.

Model Highlights

• Native MoE Scaling: Optimized for the 256-expert architecture (8+1 activated).
• Thinking Preservation: Retains reasoning context in historical messages via DeltaNet stability.

---

Acknowledgments: These quants were built using the oMLX framework. The weight optimization process is based on the GPTQ algorithm by Frantar et al.

Verified via Splats Lab Vault v2.8. These models are standard mlx-lm compatible and work with any app supporting MLX safetensors.