⚡ 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.
Carnice Qwen3.6 MoE 35B-A3B APEX GGUF
APEX (Adaptive Precision for EXpert Models) quantizations of samuelcardillo/Carnice-Qwen3.6-MoE-35B-A3B.
Brought to you by the LocalAI team | APEX Project | Technical Report
Available Files
| File | Profile | Size | Best For |
|---|---|---|---|
| Carnice-Qwen3.6-MoE-35B-A3B-APEX-I-Balanced.gguf | I-Balanced | 24 GB | Best overall quality/size ratio |
| Carnice-Qwen3.6-MoE-35B-A3B-APEX-I-Quality.gguf | I-Quality | 22 GB | Highest quality with imatrix |
| Carnice-Qwen3.6-MoE-35B-A3B-APEX-Quality.gguf | Quality | 22 GB | Highest quality standard |
| Carnice-Qwen3.6-MoE-35B-A3B-APEX-Balanced.gguf | Balanced | 24 GB | General purpose |
| Carnice-Qwen3.6-MoE-35B-A3B-APEX-I-Compact.gguf | I-Compact | 17 GB | Consumer GPUs, best quality/size |
| Carnice-Qwen3.6-MoE-35B-A3B-APEX-Compact.gguf | Compact | 17 GB | Consumer GPUs |
| Carnice-Qwen3.6-MoE-35B-A3B-APEX-I-Mini.gguf | I-Mini | 14 GB | Smallest viable, fastest inference |
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: Carnice Qwen3.6 MoE 35B-A3B (fine-tuned for agentic/tool-calling)
- Base: Qwen 3.6 35B-A3B
- Layers: 40
- Experts: 256 routed + shared (8 active per token)
- Total Parameters: ~35B
- Active Parameters: ~3B per token
- Attention: Hybrid (full attention every 4th layer, linear/Mamba otherwise)
- APEX Config: 5+5 symmetric edge gradient across 40 layers
- Calibration: v1.3 diverse dataset (chat, code, reasoning, multilingual, tool-calling, Wikipedia)
- llama.cpp: Built with b8797
Run with LocalAI
local-ai run mudler/Carnice-Qwen3.6-MoE-35B-A3B-APEX-GGUF@Carnice-Qwen3.6-MoE-35B-A3B-APEX-I-Balanced.gguf
Credits
- Carnice fine-tune: samuelcardillo
- APEX quantization: LocalAI team
- Built on llama.cpp