4-bit Open on Hugging Face
Parameters
2.7B
Weights (est.)
≈ 1.5 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/LFM2.5-2.6B-OptiQ-4bit --prompt "Hello"
Details
- Base model
- LiquidAI/LFM2.5-2.6B (quantized)
- Vendor
- LiquidAI
- Task
- Text generation
- Quantisation
- 4-bit
- Licence
- lfm1.0
- Downloads
- 330 in the last 30 days · 1.4K all time · 2 likes
- Published
- 2026-08-06 · updated 2026-10-14 (27d ago)
- Tags
- lfm2, optiq, quantized, 4bit, mixed-precision, liquid-ai, apple-silicon, on-device
Other MLX builds of LiquidAI/LFM2.5-2.6B
| Build | Quantisation | Weights | Mac memory | Downloads |
|---|---|---|---|---|
| LFM2.5-2.6B-4bit | 4-bit | ≈ 1.5 GB | 8 GB+ | 1.8K |
| LFM2.5-2.6B-OptiQ-4bit (this one) | 4-bit | ≈ 1.5 GB | 8 GB+ | 330 |
| LFM2.5-2.6B-mxfp4 | MXFP4 | ≈ 1.5 GB | 8 GB+ | 101 |
| LFM2.5-2.6B-nvfp4 | 4-bit | ≈ 1.5 GB | 8 GB+ | 92 |
| LFM2.5-2.6B-5bit | 5-bit | ≈ 1.9 GB | 8 GB+ | 103 |
| LFM2.5-2.6B-6bit | 6-bit | ≈ 2.2 GB | 8 GB+ | 124 |
| LFM2.5-2.6B-8bit | 8-bit | ≈ 2.9 GB | 8 GB+ | 220 |
| LFM2.5-2.6B-mxfp8 | MXFP8 | ≈ 2.9 GB | 8 GB+ | 92 |
| LFM2.5-2.6B-bf16 | bf16 | ≈ 5.4 GB | 16 GB+ | 64 |