---
license: apache-2.0
base_model: Jackrong/Qwen3.5-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled
tags:
  - gguf
  - quantized
  - apex
  - moe
  - mixture-of-experts
  - qwen3.5
  - claude-distilled
---

# Qwen3.5-35B-A3B Claude-Distilled APEX GGUF

**APEX (Adaptive Precision for EXpert Models)** quantizations of [Qwen3.5-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled](https://huggingface.co/Jackrong/Qwen3.5-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled).

**Brought to you by the [LocalAI](https://github.com/mudler/LocalAI) team** | [APEX Project](https://github.com/mudler/apex-quant) | [Technical Report](https://github.com/mudler/apex-quant/blob/main/paper/APEX_Technical_Report.pdf)

## Benchmark Results

Benchmarks coming soon. For reference APEX benchmarks on the same Qwen3.5-MoE architecture, see [mudler/Qwen3.5-35B-A3B-APEX-GGUF](https://huggingface.co/mudler/Qwen3.5-35B-A3B-APEX-GGUF).

## Available Files

| File | Profile | Size | Best For |
|------|---------|------|----------|
| Qwen3.5-35B-A3B-Claude-Distilled-APEX-I-Balanced.gguf | I-Balanced | ~24 GB | Best overall quality/size ratio |
| Qwen3.5-35B-A3B-Claude-Distilled-APEX-I-Quality.gguf | I-Quality | ~22 GB | Highest quality with imatrix |
| Qwen3.5-35B-A3B-Claude-Distilled-APEX-Quality.gguf | Quality | ~22 GB | Highest quality standard |
| Qwen3.5-35B-A3B-Claude-Distilled-APEX-Balanced.gguf | Balanced | ~24 GB | General purpose |
| Qwen3.5-35B-A3B-Claude-Distilled-APEX-I-Compact.gguf | I-Compact | ~17 GB | Consumer GPUs, best quality/size |
| Qwen3.5-35B-A3B-Claude-Distilled-APEX-Compact.gguf | Compact | ~17 GB | Consumer GPUs |
| Qwen3.5-35B-A3B-Claude-Distilled-APEX-I-Mini.gguf | I-Mini | ~13 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](https://github.com/mudler/apex-quant) for full details, technical report, and scripts.

## Architecture

- **Model**: Qwen3.5-35B-A3B-Claude-Distilled (Qwen3.5-MoE, distilled from Claude 4.6 Opus reasoning)
- **Layers**: 40
- **Experts**: 256 routed + 1 shared (8 active per token)
- **Total Parameters**: ~35B
- **Active Parameters**: ~3B per token
- **APEX Config**: 5+5 symmetric edge gradient across 40 layers
- **Calibration**: v1.3 diverse dataset (chat, code, reasoning, multilingual, tool-calling, Wikipedia)

## Run with LocalAI

```bash
local-ai run mudler/Qwen3.5-35B-A3B-Claude-Distilled-APEX-GGUF@Qwen3.5-35B-A3B-Claude-Distilled-APEX-I-Balanced.gguf
```

## Credits

APEX is brought to you by the [LocalAI](https://github.com/mudler/LocalAI) team. Developed through human-driven, AI-assisted research. Built on [llama.cpp](https://github.com/ggerganov/llama.cpp).