3-bit Open on Hugging Face
Parameters
8.3B
Weights (est.)
≈ 3.6 GB
Mac memory (est.)
8 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/LFM2-8B-A1B-3bit-MLX --prompt "Hello"
Details
- Base model
- LiquidAI/LFM2-8B-A1B (quantized)
- Vendor
- LiquidAI
- Task
- Text generation
- Quantisation
- 3-bit
- Licence
- lfm1.0
- Downloads
- 85 in the last 30 days · 2.1K all time · 2 likes
- Published
- 2025-10-08 · updated 2025-10-08 (1y ago)
- Languages
- en
- Tags
- lfm2_moe, apple-silicon, liquidai, lfm2, transformer, long-context, instruct, quantized, 3bit, coding, custom_code
Other MLX builds of LiquidAI/LFM2-8B-A1B
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| LFM2-8B-A1B-3bit-MLX (this one) | 3-bit | ≈ 3.6 GB | 8 GB+ | 85 |
| LFM2-8B-A1B-4bit | 4-bit | ≈ 4.7 GB | 8 GB+ | 577 |
| LFM2-8B-A1B-5bit | 5-bit | ≈ 5.7 GB | 16 GB+ | 46 |
| LFM2-8B-A1B-6bit-MLX | 6-bit | ≈ 6.8 GB | 16 GB+ | 62 |
| LFM2-8B-A1B-6bit | 6-bit | ≈ 6.8 GB | 16 GB+ | 35 |
| LFM2-8B-A1B-8bit-MLX | 8-bit | ≈ 8.9 GB | 16 GB+ | 63 |
| LFM2-8B-A1B-8bit | 8-bit | ≈ 8.9 GB | 16 GB+ | 60 |
| LFM2-8B-A1B-fp16 | fp16 | ≈ 33.4 GB | 48 GB+ | 40 |