4-bit Open on Hugging Face
Parameters
180B
Weights (est.)
≈ 102 GB
Mac memory (est.)
128 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-vlm on an Apple silicon Mac:
pip install mlx-vlmpython -m mlx_vlm.generate --model mlx-community/Qwen3.8-Flash-Next-4bit-mtp --image photo.jpg --prompt "Describe this image"
Details
- Base model
- Qwen/Qwen3.8-Flash-Next (quantized) · Qwen3.8 Flash Next on the radar →
- Vendor
- Qwen
- Task
- Vision + text
- Quantisation
- 4-bit
- Licence
- —
- Downloads
- 2.3K in the last 30 days · 4.6K all time · 6 likes
- Published
- 2026-09-23 · updated 2026-09-25 (1mo ago)
- Languages
- en
- Tags
- qwen4_exp, speculative-decoding
Other MLX builds of Qwen/Qwen3.8-Flash-Next
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| Qwen3.8-Flash-Next-OptiQ-2bit | 2-bit | ≈ 56.3 GB | 96 GB+ | 2.5K |
| Qwen3.8-Flash-Next-4bit | 4-bit | ≈ 100 GB | 128 GB+ | 7.8K |
| Qwen3.8-Flash-Next-4bit-mtp (this one) | 4-bit | ≈ 102 GB | 128 GB+ | 2.3K |
| Qwen3.8-Flash-Next-oQ6e-mtp | 6-bit | ≈ 147 GB | 192 GB+ | 891 |
| Qwen3.8-Flash-Next-oQ8e-mtp | 8-bit | ≈ 192 GB | 256 GB+ | 1.0K |