4-bit Open on Hugging Face
Parameters
35B
Weights (est.)
≈ 20.4 GB
Mac memory (est.)
32 GB+
Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.
Run it
With mlx-lm on an Apple silicon Mac:
pip install mlx-lmmlx_lm.generate --model mlx-community/Qwen3.6-35B-A3B-OptiQ-4bit --prompt "Hello"
Details
- Base model
- Qwen/Qwen3.6-35B-A3B (quantized) · Qwen3.6 35B A3B on the radar →
- Vendor
- Qwen
- Task
- Text generation
- Quantisation
- 4-bit
- Licence
- apache-2.0
- Downloads
- 3.6K in the last 30 days · 59K all time · 82 likes
- Published
- 2026-04-26 · updated 2026-11-05 (6d ago)
- Tags
- qwen3_5_moe, quantized, mixed-precision, 4bit, 8bit, optiq, apple-silicon, qwen3.6
Other MLX builds of Qwen/Qwen3.6-35B-A3B
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| Qwen3.6-35B-A3B-MTP-4bit | 4-bit | ≈ 475 MB | 8 GB+ | 1.3K |
| Qwen3.6-35B-A3B-MTP-5bit | 5-bit | ≈ 581 MB | 8 GB+ | 278 |
| Qwen3.6-35B-A3B-MTP-bf16 | bf16 | ≈ 1.7 GB | 8 GB+ | 203 |
| Huihui-Qwen3.6-35B-A3B-abliterated-4.4bit-msq | 4-bit | ≈ 5.1 GB | 8 GB+ | 282 |
| Qwen3.6-35B-A3B-4bit | 4-bit | ≈ 20.4 GB | 32 GB+ | 38K |
| Qwen3.6-35B-A3B-OptiQ-4bit (this one) | 4-bit | ≈ 20.4 GB | 32 GB+ | 3.6K |
| Qwen3.6-35B-A3B-nvfp4 | 4-bit | ≈ 20.4 GB | 32 GB+ | 1.9K |
| Qwen3.6-35B-A3B-mxfp4 | MXFP4 | ≈ 20.4 GB | 32 GB+ | 261 |
| Qwen3.6-35B-A3B-5bit | 5-bit | ≈ 24.7 GB | 32 GB+ | 3.3K |
| Qwen3.6-35B-A3B-6bit | 6-bit | ≈ 29.1 GB | 48 GB+ | 2.6K |
| Qwen3.6-35B-A3B-8bit | 8-bit | ≈ 37.7 GB | 48 GB+ | 5.6K |
| Qwen3.6-35B-A3B-mxfp8 | MXFP8 | ≈ 37.7 GB | 48 GB+ | 1.6K |
| Qwen3.6-35B-A3B-bf16 | bf16 | ≈ 70.2 GB | 96 GB+ | 2.7K |