⚡ Each donation = another big MoE quantized
I host 25+ free APEX MoE quantizations as independent research. My only local hardware is an NVIDIA DGX Spark (122 GB unified memory), enough for ~30-50B-class MoEs, but bigger ones (200B+) require rented compute on H100/H200/Blackwell, typically $20-100 per quant.
If APEX quants are useful to you, your support directly funds those bigger runs.
🎉 Patreon (Monthly) | ☕ Buy Me a Coffee | ⭐ GitHub Sponsors
💚 Big thanks to Hugging Face for generously donating additional storage, much appreciated.
Qwen3.5-397B-A17B APEX GGUF
APEX (Adaptive Precision for EXpert Models) quantizations of Qwen3.5-397B-A17B.
Brought to you by the LocalAI team | APEX Project | Technical Report
Benchmark Results
Benchmarks coming soon. For reference APEX benchmarks on the Qwen3.5-35B-A3B architecture, see mudler/Qwen3.5-35B-A3B-APEX-GGUF.
What is APEX?
APEX is a quantization strategy for Mixture-of-Experts (MoE) models. It classifies tensors by role (routed expert, shared expert, attention) and applies a layer-wise precision gradient -- edge layers get higher precision, middle layers get more aggressive compression. I-variants use diverse imatrix calibration (chat, code, reasoning, tool-calling, agentic traces, Wikipedia).
See the APEX project for full details, technical report, and scripts.
Architecture
- Model: Qwen3.5-397B-A17B (qwen3_5_moe)
- Layers: 60 (hybrid: linear attention + full attention every 4th layer)
- Experts: 512 routed (10 active per token)
- Total Parameters: ~397B
- Active Parameters: ~17B per token
- Vision: Built-in vision encoder (mmproj included)
- Context: 262K tokens
- APEX Config: 5+5 symmetric edge gradient across 60 layers
- Calibration: v1.3 diverse dataset (chat, code, reasoning, multilingual, tool-calling, Wikipedia)
Run with LocalAI
local-ai run mudler/Qwen3.5-397B-A17B-APEX-GGUF@Qwen3.5-397B-A17B-APEX-I-Balanced.gguf
Credits
APEX is brought to you by the LocalAI team. Developed through human-driven, AI-assisted research. Built on llama.cpp.